Thanks for a great article that sheds light on one aspect of DNA trait research. Everything about DNA research is complex. I follow many facets of human DNA research. A couple of my favorites are psychometric analyses of human traits with an effort to determine genetic causes and human genetic enhancement. With a little irony, AI has already become a valuable tool in human genetic research, and someday very well may lead to the enhancement of human intelligence.
Did you reach out to Roisman & Fraley for their reactions to your critique? They are heavy hitters in the field and I would be curious to see how they respond to your points. Your third point still remains a bit unclear to me as the remaining 5% would be a mixture of non/shared + E?
I didn't. But if they want to answer the question of genetic confounding, it could be done on the cheap with an adoption study (many datasets), a children-of-twins design (registry data), or a similar design involving cousins (registry data).
Their design didn't address the question. That doesn't mean the association is actually due to genetic confounding; it could still be an environmental effect, we just don't know.
Thanks for a great article that sheds light on one aspect of DNA trait research. Everything about DNA research is complex. I follow many facets of human DNA research. A couple of my favorites are psychometric analyses of human traits with an effort to determine genetic causes and human genetic enhancement. With a little irony, AI has already become a valuable tool in human genetic research, and someday very well may lead to the enhancement of human intelligence.
HBD to Sir Francis Galton! 🥳🎂🔔
Did you reach out to Roisman & Fraley for their reactions to your critique? They are heavy hitters in the field and I would be curious to see how they respond to your points. Your third point still remains a bit unclear to me as the remaining 5% would be a mixture of non/shared + E?
I didn't. But if they want to answer the question of genetic confounding, it could be done on the cheap with an adoption study (many datasets), a children-of-twins design (registry data), or a similar design involving cousins (registry data).
Thanks. I was able to discuss your critique a bit with Roisman.
Their results were so unlikely even at a glance that I knew something must be up, and that they were trying to arrive at a desired result.
Their design didn't address the question. That doesn't mean the association is actually due to genetic confounding; it could still be an environmental effect, we just don't know.