Hi Hunter,
Thank you for your well-diagramed and thoughtful DAG. It is helpful to visualize your causal pathway as well as the potential confounders/mediators. I think one of the challenging aspects about figuring out if SES is a confounder, effects modifier, or mediator is how we define SES. As we have learned, there are many ways by which try to account for the impact of SES on our outcomes and there are many ways that SES can be measured and used. I think this leads to a lot of heterogeneity in defining SES and makes it challenging to figure out if SES as defined by each researcher is a confounder, effects modifier, or mediator. I agree that it depends on how the investigator team decides the DAG is drawn. I would hope that rather than going on a fishing expedition, that we use the research currently available in order to come up with a plausible DAG and define the role of SES based on prior evidence. Of course as more research is done and statistical analyses techniques become more sophisticated, we will be able to better account for the various ways that SES impacts health outcomes on individual, community, societal levels, including the interaction between those different levels.
Thank you for sharing your thoughts!
Janet
Thank you for your well-diagramed and thoughtful DAG. It is helpful to visualize your causal pathway as well as the potential confounders/mediators. I think one of the challenging aspects about figuring out if SES is a confounder, effects modifier, or mediator is how we define SES. As we have learned, there are many ways by which try to account for the impact of SES on our outcomes and there are many ways that SES can be measured and used. I think this leads to a lot of heterogeneity in defining SES and makes it challenging to figure out if SES as defined by each researcher is a confounder, effects modifier, or mediator. I agree that it depends on how the investigator team decides the DAG is drawn. I would hope that rather than going on a fishing expedition, that we use the research currently available in order to come up with a plausible DAG and define the role of SES based on prior evidence. Of course as more research is done and statistical analyses techniques become more sophisticated, we will be able to better account for the various ways that SES impacts health outcomes on individual, community, societal levels, including the interaction between those different levels.
Thank you for sharing your thoughts!
Janet