Comment About the study... (Score 1) 81
I haven't posted to Slashdot in a long time, but thought that since this is about something near to me, which doesn't happen often, I would.
I'm loosely involved with Avshalom Caspi's group (Caspi is the lead author of the study, and happens to be at U of Wisconsin as well as UC London). I don't work directly with Avshalom on a daily basis, but I know him, and work with their data, etc. on a fairly regular basis. I wasn't involved in this particular study, but knew about it ahead of time.
Anyway, there's a couple of things to keep in mind about this study before you take it too much to heart. First, there's no replication sample in this study. While that isn't always an issue in science, with behavioral genomic studies it is, because effect sizes are typically so small to begin with, and false positives are an ongoing concern. It's of particular concern in this study because they're claiming an interaction effect. Now, you'd think that interactive effects would be the norm in psychology, but in fact, they're notoriously difficult to replicate, and almost never do. Not never, but rarely. If this effect didn't reliably replicate, I wouldn't be surprised in the least.
Second, there's no background genetic controls. If MAO polymorphisms had never been included in the study, Avshalom would have been laughed off the scene because of the lack of genetic controls (e.g., using twin, adoption, or family designs). There's a lot of genetic variance in that sample that is related to externalizing behavior that's not being accounted for by the single gene that they considered in the study. It's entirely possible that that background genetic variation is accounting for the interaction with MAO polymorphisms and not abuse per se.
It's a great study, with a lot to think about, and I think more people should do more studies like it. But I am a bit reluctant to make too much of it at his point, because there's a lot of things that were missing in the study, things that were especially important given the complexity of the claims they were making.
I'm loosely involved with Avshalom Caspi's group (Caspi is the lead author of the study, and happens to be at U of Wisconsin as well as UC London). I don't work directly with Avshalom on a daily basis, but I know him, and work with their data, etc. on a fairly regular basis. I wasn't involved in this particular study, but knew about it ahead of time.
Anyway, there's a couple of things to keep in mind about this study before you take it too much to heart. First, there's no replication sample in this study. While that isn't always an issue in science, with behavioral genomic studies it is, because effect sizes are typically so small to begin with, and false positives are an ongoing concern. It's of particular concern in this study because they're claiming an interaction effect. Now, you'd think that interactive effects would be the norm in psychology, but in fact, they're notoriously difficult to replicate, and almost never do. Not never, but rarely. If this effect didn't reliably replicate, I wouldn't be surprised in the least.
Second, there's no background genetic controls. If MAO polymorphisms had never been included in the study, Avshalom would have been laughed off the scene because of the lack of genetic controls (e.g., using twin, adoption, or family designs). There's a lot of genetic variance in that sample that is related to externalizing behavior that's not being accounted for by the single gene that they considered in the study. It's entirely possible that that background genetic variation is accounting for the interaction with MAO polymorphisms and not abuse per se.
It's a great study, with a lot to think about, and I think more people should do more studies like it. But I am a bit reluctant to make too much of it at his point, because there's a lot of things that were missing in the study, things that were especially important given the complexity of the claims they were making.
