The Association

February 26, 2009
Posted by Jay Livingston
researchers found that teenagers who preferred popular songs with degrading sexual references were more likely to engage in intercourse or in pre-coital activities.
That’s from the lead of an article passed around on a campus list here. America’s teens are having sex, and music is taking the rap.

I don’t know how far back in history this idea goes – blaming sex on music – maybe to the ancient Athenians. In the past century in the US, we’ve gone from ragtime to rap, each generation worried that the raunchiness of the music and lyrics their kids were listening to was leading those kids to sinful pleasures.

If each generation was right about the increase over the previous generation in a continually upward curve, kids today would have run out of hours in the day to have sex or “pre-coital activities” (just which base are they talking about anyway, and why didn’t we ever have an equivalent of shortstop?).

Now we have Research and Science to justify the fears about music. Note the clear cause-effect relation implied in that first sentence. Kids who listened to nasty music were more likely.

Here’s what the article* actually said
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
The emphasis is my own addition because somebody here is missing a point that any intro sociology student should have learned: correlation is not cause.

Back in the 60s there was a rock group called The Association. (Anybody else remember “Along Comes Mary”?) I think they chose that name to distinguish themselves from another group, The Causation.

With Association, you don’t know what’s causing what. The message of that first sentence is that listening to those terrible, horrible, no good, very bad lyrics makes kids go out and have sex. But an equally plausible explanation is that kids who like sex in real life also like it in their music.

Even if there were a time factor with exposure to the nasty music coming first, you still couldn't conclude causation. All you could say is that kids who like to listen to dirty lyrics when they're young grow up to like doing dirty things when they get a little older.

And oh, don't bother Googling for The Causation or their greatest hits. I just made that part up.



* “Exposure to Sexual Lyrics and Sexual Experience Among Urban Adolescents,” by Brian A. Primack, MD, EdM, MS, Erika L. Douglas, MS, Michael J. Fine, MD, MSc, and Madeline A. Dalton, PhD. It appears in the American Journal of Preventive Medicine, Volume 36, Issue 4 (April 2009).

3 comments:

Corey said...

Re: "The emphasis is my own addition because somebody here is missing a point that any intro sociology student should have learned: correlation is not cause..

[Sarcasm]
Yeah, but Jay... the correlation is significant at .05 level... that means that it's independently true of any criticism... right?
[/sarcasm]

Sigh... actually correlation should be used to establish evidence of causation. At least that's the way it's supposed to work, when the model has been properly specified on a well established theoretical foundation. I haven't read the article in question that establishes "independent correlation", but I would hope that the authors actually did some conceptual/theoretical work in establishing that theoretical linkage between rap and sex. My guess is that they didn't, or did so in such a thin manner that it is the equivalent of not theorizing at all.

My problem with much of modern "empirical" (and I use that term loosely) social science is that it skips the whole conceptualizing process. Step 1: Get a topic; Step 2: Get a data set; Step 3 run some math on the data set; Step 4, create hypotheses post-hoc that fit the statistically significant associations; Step 5 write it up as if the hypothesis was established on the front end.

Incidentally, Jeremy Freese wrote a very readable overview of good practice for survey analysis in his chapter on the Secondary Analysis of Large Survey Data Sets.

Jay Livingston said...

Corey, Here's the link to the post about the article http://www.physorg.com/news154683399.html. (The article itself is not yet available at the journal Website.) The title of the post is "Sexual lyrics in popular songs linked to early sexual experiences." "Linked" is one of those words you here when the speaker wants to imply a stronger connection than is probably there ("WVU Prof Linked to Organized Crime").

Corey said...

Re: "WVU Prof Linked to Organized Crime"

Damn! I've been discovered. I'm with you Jay. I'm waiting for the article to come out so I can examine the argument in more detail. Off the cuff, I'm willing to buy the "linkage" between lyrics and behavior, insofar as lyrics are a proxy for consumption behavior of certain groups of people. But if I'm right, then the linkage isn't between the lyrics... (Dr. Dre made me do it); it's between the social groups and the behavior.