Admittedly, the title of this post is kind of a straw man. Whether behavior genetic methods are appropriate for relationships research depends on what your question is.And there are lots of potential questions one could ask. To list just a few:
- Are there genetic influences on relationship outcomes?
- Do relationship factors have a causal effect on some other aspect of individual functioning (e.g., the onset of alcohol use disorder or desistance from antisocial behavior)?
- Do relationship factors moderate genetic influences on an outcome (e.g., are genetic influences on alcohol use attenuated for folks who are in relationships?)
My focus here is on considering whether the models that are most prevalent in behavior genetics research -- variance partitioning twin models -- are appropriate for relationships research. As I mentioned in my last post, these models compare patterns of resemblance for monozygotic and dizygotic twins to make inferences about the relative degree of genetic and environmental influences on an outcome.
In these types of models, each individual has a phenotype (or outcome). You're either divorced, or you're not. You're either more or less satisfied with your relationship. This kind of modeling is, in many ways, antithetical to how a relationships researcher thinks about marital stability or relationship satisfaction. That's because these outcomes don't "exist" within individuals. They are dyadically constructed, meaning that they reflect some mish-mash of partners' characteristics. And that mish-mash is probably not a linear combination of Partner 1's divorce genes + Partner 2's divorce genes.
One of my favorite illustrations of dyadic thinking is the link between commitment and relationship stability. Let's say you have two couples, and you ask everyone how committed they are to that relationship on a 1-10 scale, with 10 indicating a higher level of commitment. In Couple 1, Partner A reports a 5, and Partner B a 7. In Couple 2, both partners report a 5. Which couple will last longer? Based on previous findings, we expect that Couple 1 will break up before Couple 2. And that's because there is a greater discrepancy between partners' self-reports of commitment, even though their "average" level of commitment is higher than Couple 2. And this discrepancy between partners can be even more de-stabilizing than a low level of commitment.
This of course doesn't mean that genetic and environment variance components that you get out of a twin model for something like divorce or relationship satisfaction are "wrong". Rather, I'm trying to point out that the models that are typically used in behavior genetics don't always capture the complexity of how genetic and environmental factors come together to influence relationship outcomes. Most notably, a conventional twin model can't take into account the contributions of both partners' genes and environments.
In these types of models, each individual has a phenotype (or outcome). You're either divorced, or you're not. You're either more or less satisfied with your relationship. This kind of modeling is, in many ways, antithetical to how a relationships researcher thinks about marital stability or relationship satisfaction. That's because these outcomes don't "exist" within individuals. They are dyadically constructed, meaning that they reflect some mish-mash of partners' characteristics. And that mish-mash is probably not a linear combination of Partner 1's divorce genes + Partner 2's divorce genes.
One of my favorite illustrations of dyadic thinking is the link between commitment and relationship stability. Let's say you have two couples, and you ask everyone how committed they are to that relationship on a 1-10 scale, with 10 indicating a higher level of commitment. In Couple 1, Partner A reports a 5, and Partner B a 7. In Couple 2, both partners report a 5. Which couple will last longer? Based on previous findings, we expect that Couple 1 will break up before Couple 2. And that's because there is a greater discrepancy between partners' self-reports of commitment, even though their "average" level of commitment is higher than Couple 2. And this discrepancy between partners can be even more de-stabilizing than a low level of commitment.
This of course doesn't mean that genetic and environment variance components that you get out of a twin model for something like divorce or relationship satisfaction are "wrong". Rather, I'm trying to point out that the models that are typically used in behavior genetics don't always capture the complexity of how genetic and environmental factors come together to influence relationship outcomes. Most notably, a conventional twin model can't take into account the contributions of both partners' genes and environments.
RSS Feed