• A good paper (instrumental variables, substance use)

    It’s a nice coincidence that the very month I started working on a substance use disorder (SUD) project that will employ instrumental variables (IV) techniques Michael French and Ioana Popovici published “That instrument is lousy! In search of agreement when using instrumental variables estimation in substance use research.” The paper is chock full of useful content and references, for me anyway.

    Actually, Section 2 is a generic and complete, though technical, summary of the IV approach with references to key literature. That alone is valuable. But what about those “lousy” instruments? Well, I’m not going to discuss all of them or very much of the paper. I’ll just highlight a few things for my own future reference.*

    One genre of instruments for SUD or SUD treatment one might consider is state-level, such as state laws, taxes, policies and prices. State minimum legal drinking age might be correlated with regional attitudes toward drinking but not consequences of SUD or SUD treatment, for example. Minimum drinking age is 21 nationally, but the state-based age just prior to the national law could still reflect attitudes today.

    Likewise, the blood alcohol concentration for DUI charges, consumption or sales volumes, taxation levels on alcohol, taxation levels on cigarettes have all been used as instruments in studies of alcohol use. For these, and all other instruments, French and Popovici cite papers in which they are used. They also point out that all such regional variables could be correlated with unobservable state attitudes that are correlated with consumption and outcomes. They may not be good instruments. It depends on the application.

    Instruments for drug use described in the paper include cocaine prices, marijuana decriminalization, penalties for marijuana possession, and drug possession arrest rates.

    If one is considering an IV approach for an SUD study, this French and Popovici paper is worth consulting. In fact, it may be hard to publish the results of such a study in an economics journal without doing so, and heeding their advice (none of which I’ve reproduced in this post).

    * Lame post, I know. They can’t all be winners. This is really a “notebook entry” type post. Plus, it is Friday, we got more snow, and my left shin aches a little. Enough excuses for you?

    Comments closed