This is a good point. I was recently having a conversation with a non-scientist (I am not one either) about two different studies. One was an RCT that found no

This is a good point. I was recently having a conversation with a non-scientist (I am not one either) about two different studies. One was an RCT that found no effect for a treatment, and the other was a "case control, test-negative" design study that found a beneficial effect. My friend said the latter was as good as an RCT because it also has a control arm. But when I looked up what "case control, test-negative" means, I found that it is observational, and there is no intervention provided to any of the individuals. Some of the recent discussions of the design indicate that it is understood by scientists, but can mislead laypeople, who assume that its conclusions are much more robust than they actually are. We should definitely continue to do research of many types, partly as a way to figure out where to spend the time/money to do RCTs (which are more expensive than other types of analysis). But we need people reporting on studies to be very clear up-front when they are not describing an RCT. They should say what the study's conclusion means, versus what it would mean if there were a similar 'finding' in an RCT. Otherwise people will not understand that they are being told a weak conclusion, not realizing there is a strong conclusion that has gone unmentioned.