Do Parental Consent Laws Decrease Risky Teen Sex?
Marginal Revolution points to a colloquium paper arguing that parental consent laws serve to lower the incidence of risky sex for White and Hispanic teens. The authors, Klick and Stratmann, use gonorrhea infection rates as a proxy for risky sex. They claim, as have others, that abortion serves an insurance function and as that is lessened (via parental consent), teens respond by having less sex or less risky sex (ie, increased condom usage).
While the argument does make some intuitive sense, there is a key assumption underlying their analysis. That is, teens are aware of parental consent laws in their state. Nowhere in their design is any variable measuring knowledge or understanding of the applicable law. Nor is there any attempt to use a proxy for knowledge. While I am not suggesting that such data exist or could be teased out of some proxy, it does seriously undermine the validity of their conclusions. Could it be that some other unobservable variable is driving lower gonorrhea rates in those years? For example, we know that teen pregnancy rates vary year to year (though they do seem to follow a somewhat linear trajectory) so why should we not also assume that STD rates might also vary?
In addition, Klick and Stratmann code the state law variable 1 for full year of consent, 0 for no consent and a number between corresponding to partial year consent law. From my quick reading of their paper and model, it seems unclear that there is a way to control for year to year spillover effects. For example, what if state A has a stringent parental consent law in year X which gets eliminated in year Z. Are teens responding to what they perceive the law to be? Again, this is a knowledge of law problem. Can we really assume that teens are aware?
This paper does little to convince me that parental consent laws will affect the incidence of risky sex among teens. Of course, this leaves out the whole issue of whether parental consent laws are harmful in other ways (ie, family dynamics).
While the argument does make some intuitive sense, there is a key assumption underlying their analysis. That is, teens are aware of parental consent laws in their state. Nowhere in their design is any variable measuring knowledge or understanding of the applicable law. Nor is there any attempt to use a proxy for knowledge. While I am not suggesting that such data exist or could be teased out of some proxy, it does seriously undermine the validity of their conclusions. Could it be that some other unobservable variable is driving lower gonorrhea rates in those years? For example, we know that teen pregnancy rates vary year to year (though they do seem to follow a somewhat linear trajectory) so why should we not also assume that STD rates might also vary?
In addition, Klick and Stratmann code the state law variable 1 for full year of consent, 0 for no consent and a number between corresponding to partial year consent law. From my quick reading of their paper and model, it seems unclear that there is a way to control for year to year spillover effects. For example, what if state A has a stringent parental consent law in year X which gets eliminated in year Z. Are teens responding to what they perceive the law to be? Again, this is a knowledge of law problem. Can we really assume that teens are aware?
This paper does little to convince me that parental consent laws will affect the incidence of risky sex among teens. Of course, this leaves out the whole issue of whether parental consent laws are harmful in other ways (ie, family dynamics).
posted by Anonymous at 1:17 PM
0 Comments:
Post a Comment
<< Home