Jillaine, I hope I am not insulting anyone by saying this, but I would say that the MAJORITY of profiles I have seen are unsourced. So using "changes" as a proxy for "a real human being has actually entered data for this person", you could program the computer to destroy all links where there are not two individuals listed under changes, and voila, we would get rid of a lot of links where there is lack of evidence. This would be a blanket approach to the issue and I think it would be a terrible idea.
So I think the situation you are describing is not a blanket approach to all the links without evidence, but a particular link you are looking at, at the moment. At that moment, you do have one piece of evidence -- by no means conclusive, but evidence: The person who created the link believed it should be there. And your gut gives you another piece of evidence: "out of all the undocumented profiles on Wikitree, is there something about this one that calls for my attention?" If your gut says no, you move on. Your time is limited. If your gut says yes, you're hooked, because now you have two tiny pieces of competing evidence and you want much more. This is what makes genealogy and history addictive.
My basic point is that at that particular moment, you have too much evidence to walk away, but you don't have enough evidence to just go in and break the link, using your gut to override the original poster's belief. At that particular moment, you are now driven to find enough evidence to confirm or override the original poster's belief.
By including "your gut feeling" as evidence, I'm using the word "evidence" broadly, but I think it's important to do so. We come across profiles that don't "pass the smell test" but it takes further exploration to figure out what's wrong. Sometimes it may be obvious. A woman having a child at age 5, or a couple marrying across a space of a hundred years, are not troubled by "lack of evidence" -- those narratives are troubled by the presence of the competing evidence of human biology and experience.
On Magna Carta, I had to break a parental link that a lot of people believe in and people have written books to support. Richardson took the short route -- Giles Brent's will identified 3 children, therefore he had 3 only children. I felt compelled to take the longer route and research -- and shoot down -- the competing theories that Giles had additional children whom he had some reason for not naming in his will.
Magna Carta starts out with an unusually strong beginning -- well researched documentation/evidence for a whole line of people from the Magna Carta Surety down to the Gateway Ancestor. So with the lines we're documenting, there is no such thing as lack of evidence -- only, sometimes, lack of someone having put the evidence on WikiTree. And we're fixing that.
In the end, it really is all about the evidence. Often the starting point is that it's not there. But that's the starting point. For me, the solution is to get the evidence, not to just disconnect people that somebody believes should be together but hasn't documented. Trust me, I will be disconnecting people in the future. But I feel I need to be able to explain why with evidence of my own sufficient to override the evidence of the original poster's belief.