@jumbanho well a RCT isn't necessarily a direct measure of causality. Just because you change this and see that result doesn't mean this change caused that result.

But I'm also not sure why you're so interested in bringing in less direct indicators, as if being farther from the situation necessarily improves the observation.