1

I do enjoy these factoids, but I'm not a fan of correlation studies. This one seems to posit that sex or drugs causes depression, but does it account for pre-existing depression? A study can not determine causation.

It seems more likely to me that these teens were already depressed and tried to assuage their condition with base pleasures.

2

kbiel: Considering that the title of the article cited is "Which Comes First in Adolescence?--Sex and Drugs or Depression?", the authors presumably are addressing precisely your question....

3

It's all George Bush's fault.

4

kbiel,

I would argue a study can determine or suggest causation if done properly.

The information at familyfacts.org does not go in depth, but it does say

"Sample or Data Description
Waves I (collected in 1995) and II (1996) of the contractual data set of Add Health. Wave I included 18,924 respondents. Wave II included 13,570 respondents. Final sample for study included 13,491 youth were in 7th to 11th grade at Wave I and 8th to 12th grade at Wave II. "

So it appears they asked the respondents about their behaviors and feelings of depression, then followed up with the same group a year later. The likelihood of a person that was not depressed in the first survey, but was depressed in the second survey, was impacted by whether or not they abstained. In other words, it is suggesting causation.

Based heavily on the assumption the followed up with the same group; if not, then the survey probably isn’t worth the time I spent speculating…

5

Paul,

those with multiple sex partners were ... more likely ... to become depressed and those who used drugs frequently were about seven and a half times as likely as peers who did not use drugs to become depressed.

Gee, I don't know. The title not withstanding, that statement seems to assign causation, not question it.

I would argue a study can determine or suggest causation if done properly.

You might argue that, but you could only be right about the latter item. A correlation study might, might suggest a starting point for researchers to begin their search for causation. On the other hand, it could lead them on a wild goose chase. A correlation study can never prove causation as by definition it does not look at the mechanics of the situation, only the data surrounding it.

Correlation studies can be helpful and are sometimes useful in making a point, but they are out of line when trying to assign causation. And any self-reporting study is suspect in my mind. Perhaps something critical happened that changed their understanding of the question "are you depressed" between surveys. For example, maybe a majority of them took a health class between surveys and decided after learning about depression that they were depressed. Those that answered positive to drug use and promiscuity in the first survey, but no to depression now have a new self diagnosis that may have nothing to do with the reality of when they became depressed.

Now, had authors only stated that those students who used drugs and/or were promiscuous were often depressed, that would be an acceptable conclusion to draw from this data (ignoring the self-reporting aspect of it). It also might be more acceptable if there were a clinical diagnosis of depression or non-depression (which itself is fraught with questions) during each survey period.