Hundreds of newspapers and websites around the country have been buzzing with the news that sexually degrading song lyrics are linked to early sexual activity. The newspapers slavishly copied the press release issued by the American Journal of Preventive Medicine:

… Brian A. Primack, MD…, states, "This study demonstrates that, among this sample of young adolescents, high exposure to lyrics describing degrading sex in popular music was independently associated with higher levels of sexual behavior. In fact, exposure to lyrics describing degrading sex was one of the strongest associations with sexual activity...These results provide further support for the need for additional research and educational intervention in this area."

Surveys were completed by 711 ninth-grade students at three large urban high schools. These participants were exposed to over 14 hours each week of lyrics describing degrading sex. About one third had previously been sexually active. Compared to those with the least exposure to lyrics describing degrading sex, those with the most exposure were more than twice as likely to have had sexual intercourse…

Similarly, among those who had not had sexual intercourse, those in the highest third of exposure to lyrics describing degrading sex were nearly twice as likely to have progressed along a noncoital sexual continuum compared to those in the lowest third…

There are a lot of problems with this study, but the primary problem is not visible to lay people: the press release was issued well in advance of the publication of the article. In fact, the article is not available until the April issue of the journal is published. Journalists can request an advance copy of the article, but judging by the newspaper articles, most simply copied the press release. That means that the public has no way to independently assess the validity of the study or to determine if the author’s conclusions are justified by the data in the paper.

On its face, the idea behind the study is perfectly reasonable. Early sexual activity among teens leads to significant public health problems like the epidemic of sexually transmitted diseases and teen pregnancy. Identifying factors that promote early sexual activity might be helpful in creating public education programs to prevent it. It is entirely possible that sexually suggestive media promote early sexual activity. Unfortunately, the methodology of this study renders its conclusions highly doubtful.

In the study, Exposure to Sexual Lyrics and Sexual Experience Among Urban Adolescents (no link available), the authors recruited over 700 high school students with varying amounts of pre-existing sexual experience. The authors “inferred” the amount of sexually suggestive lyrics to which each teen was exposed by a convoluted algorithm:

They asked each teen to identify his favorite music artist and how many hours per week the teen spent listening to all music. Then they assumed that the teen’s favorite artist was representative of all music that they heard during the week. Further, they assumed that proportion of that artist’s use of sexually degrading music across his entire repertoire was representative of the teen’s exposure to sexually degrading lyrics.

For example, if a teen reported listening to 21 hours of music a week, and his favorite artist used sexually degrading lyrics in 67% of his songs, it was assumed that the teen listened to 14 hours of sexually degrading lyrics per week. As the authors acknowledge, their method has tremendous potential for error:

 It is possible that an adolescent’s favorite artist does not represent total exposure, either because the adolescent lacks access to that artist’s music (for financial or other reasons) or because the adolescent listens to a wide variety of music…

 In truth, the authors had literally no idea how many hours each teen was exposed to sexually degrading lyrics.

Moreover, the choice of favorite artist who uses sexually degrading lyrics may reflect the teen’s personal values. Those who are sexually active may more likely to enjoy sexually degrading lyrics.

The authors present the results of the study:

Participants were exposed to an average of 31.0 hours of popular music per week (SD=21.1) and an estimated 14.7 hours each week of songs with lyrics describing degrading sex (SD=17.0). Thirty percent of the 711 participants had previously had sexual intercourse…

Compared to those in the lowest tertile, those with the most exposure to lyrics describing degrading sex had greater odds for having had sexual intercourse (OR_2.07; 95% CI_1.26, 3.41). Likewise, compared to those in the lowest tertile, those with the most exposure to lyrics describing degrading sex had greater odds for having progressed further down the noncoital sexual continuum (OR_1.88; 95% CI_1.23, 2.88).

In other words, high exposure to sexually degrading lyrics was associated with a doubling of sexual activity. That sounds impressive until you consider that there were many other factors that were also associated with doubling of sexual activity. Age greater than 15 and rebelliousness doubled the level of sexual activity, and being black more than doubled the level of sexual activity.

The authors suggest that the association between exposure to sexually degrading lyrics and increased sexual activity is important and might even indicate that increased exposure to sexually degrading lyrics leads to increased sexual activity. However, a fundamental rule of scientific analysis is that correlation does not equal causation. Just because something is associated with a change does not mean that it caused the change.

The authors’ data demonstrates the importance of this rule. While age greater than 15 was associated with double the level of sexual activity, it would be absurd to suggest that increased age causes increased sexual activity. Although black teens reported double the level of sexual activity compared to white teens, it would be absurd to propose that black race causes increased sexual activity. Similarly, it is an unjustified stretch to suggest that increased exposure to sexually degrading lyrics causes increased sexual activity.

The authors openly acknowledge that their method of calculating exposure for sexually degrading lyrics has potential for large error and that correlation does not equal causation, and that causation can go both ways, but then they throw caution to the wind:

In summary, adolescents are heavily exposed to lyrics describing degrading sex in popular music, and this exposure is associated with early sexual experience among them in an urban population of youth at high risk for risky sexual behavior. These results provide further support for the need for additional research and educational intervention in this area.

But that’s not what the paper showed. We have no idea of the actual exposure of adolescents to sexual degrading lyrics. While exposure appears to be associated with early sexually experience, the three other factors that are associated with early sexual experience are clearly not causative. Even if there is a causative relationship, it is just as like to be that sexually active teens prefer sexually degrading lyrics, not that sexually degrading lyrics lead to increased sexual activity.

Finally, it was improper for the authors and the journal to send out press releases far in advance of publication of the paper. The press release is supposed to highlight the reasons to read the paper, not substitute for the paper itself. Deliberately sending out the press release in the absence of the paper deprives journalists and the public from the opportunity of actually evaluating the paper. It forces journalists to publicize the authors’ conclusions instead of their own conclusions. Maybe that was the point all along.