In my search for empirically based literature on hip-hop I found an article that deals with the change in content of hip-hop lyrics over time. I’ve wanted to do something like this, and started a bit of it a while back, but this is the first article I’ve seen that has been fairly rigorous about it. The abstract:

This paper explores the role of changing images of drinking and alcoholic beverage use in rap music from its beginnings in the United States in the late 1970s to the late 1990s. A sample of 341 rap music song lyrics released from 1979 to 1997 were selected using Billboard and Gavin rating charts. Song lyrics were coded for music genres, alcohol beverage types and brand names, drinking behaviors, drinking contexts, intoxication, attitudes towards alcohol and consequences of drinking. From 1979 to 1997, songs with references to alcohol increased fivefold (from 8 to 44%); those exhibiting positive attitudes rose from 43% to 73%; and brand name mentions increased from 46% to 71%. There were also significant increases in songs mentioning champagne and liquor (mainly expensive brand names) when comparing songs released after 1994 with those from previous years. In addition, there were significant increases in references to alcohol to signify glamour and wealth, and using alcohol with drugs and for recreational purposes. The findings also showed that alcohol use in rap music was much more likely to result in positive than negative consequences. Many of these findings are consistent with the idea that rap music has been profoundly affected by commercial forces and the marketing of alcoholic beverages. In addition, it is possible that the increase in references to alcoholic beverages in rap music, particularly spirits, is a reflection of a broader advertising culture which increasingly associates African Americans with alcohol use.

A few things jump out here. The first is the combined use of Gavin ratings and Billboard. If you want to be systematic about lyrical analysis you’ve got to figure out how you are choosing the tracks. You can’t choose them all right…or rather you CAN, but you know you’re going to get a lot of tracks that had no impact whatsoever on the consuming public. You want tracks that are both somehow representative, and tracks that were consumed by a broad range of the listening public. Pick the top fifteen, and use two metrics rather than one, and you’re on the right path.

The second thing is that the author actually did some quantitative analysis on the change in lyrical content over time. Could be that the increase between year A and year B as far as presence of alcohol related lyrics goes, is totally due to chance.

The third is that she not only coded the presence of alcohol-related content in the tracks, but she also coded what they were in relation to (did it pop up because it was a convenient rhyming word, or did it pop up when the MC was talking about being at the club?) and whether it was positive or negative.