References are not automatically required in any other StackExchange site, not even the ones that are much more based on rigorous scientific research. One adds them when necessary and appropriate. A blanket rule to always include references is counterproductive, especially in a soft-science topic like health which suffers from a lack of rigorous scientific research, where you sometimes have not well motivated research being done done and bad conclusions drawn as a result.One should not confuse "health" with "medicine" where one does try to uphold more rigorous standards but where things are still far from perfect compared to e.g. physics.
Simple example: can exposure to microwave radiation from mobile phones affect health? One can look at the peer reviewed literature published in medical journals, the conclusions are all over the place and you then have a WHO report that out of an abundance of caution argues caution. But simple physics will tell you that negative health effects are pretty much ruled out to a degree that you can't extract from the medical literature. That fact can, of course, also be extracted from the scientific literature, but the point I'm making here is that there is hardly any value of including references in an otherwise well written answer. You can pretty much get a citation for pretty much any claim you write down on this topic within a huge margin, much larger than what is plausible given all the scientific knowledge.
When it comes to general health issues that has more to do with how the human body functions or is supposed to function when it is not ill as opposed to problems caused by illness, then problems are much larger. If you have developed a drug X to treat illness Y, then the burden of proof is on you to prove that X works to cure Y. That's a straightforward problem we know what the end point is so the research you're going to do is well defined.
But if the problem is whether vitamin D is good for health, then you don't know a priori what to look for which severely complicates things. In case of drug X which is not made by the human body nor found in a normal diet, the null hypothesis can justifiably be taken to be that it does nothing, so that the burden of proof is fully on you to show that it works. In case of vitamin D we know that it does have a biological role also beyond bone health. So, taking the null hypothesis to be that it does nothing like in case of drug X would likely lead to skewed result. It may still be useful in the sense of showing what some RCT have proven but I would never stop taking my high dose vitamin D pills just because a RCT did not find a link between vitamin D and heart disease.
Nevertheless, the official advice in the medical literature is based on treating vitamin D just like that drug X , which is why I don't take it serious. I can of course still motivate what I'm doing from citing the literature but the reason why I would then cite these sources is then based on my own reasoning, and that's why I think it would not add much to include those refs. I could still include them to refer to refer readers to more information, but that should not per se be construed as additional support for the validity of my reasoning. Whether I then decide to add the refs or not is then purely an editorial decision up to me, it should not be a requirement to validate the answer.
