In my opinion, as a biologist, answers should include reference to only the primary scientific literature (refereed journals), scholarly texts, and textbooks. As much as possible, I include only these when I answer.
Wikipedia can be useful as a quick reference but it has consistency and accuracy problems. One only need compare the entries for Cyclostomata and Agnatha to see the information conflict. For example, the Cyclostomata entry begins by saying the article is about the class of jawless fish but the "Scientific Classification" sidebar calls it a Superclass, with two classes. The entry for Agnatha says it is a superclass for the jawless fishes. The sidebar does not mention which classes it includes. I've found many similar issues with it, which is why I do not allow my students to use it as a reference. However, you can sometimes find useful links to the primary literature, although many citations in Wikipedia are not to the primary literature.
News sites like BBC and CNN (to pick two at random) are often heavily filtered and edited by people without scientific backgrounds so the information loses most of its value. This is even worse for pop news sites like Buzzfeed, etc, or sites like Yahoo answers. Plus, I daresay that the first reference used by many of these authors is Wikipedia.
Many scientific blogs are excellent sources but I would only use them if I know the blog author is a scientist (e.g., Why Evolution is True by Jerry Coyne or John Hawks' Anthropology site) and the blog contains sufficient detail to answer the question. The blogs I tend to follow are written by scientists that have a "way with words." They can break down the scientific literature for the lay person better than I. Even so, I would try to follow through to the original literature if possible.
I know my approach has limitations from the perspective of a "general" Internet user, whatever that may be. First, not all users will have access to the literature I cite. To circumvent this as much as possible, I try to link to freely available articles (open access, author's website) or at least well-written abstracts that contain the primary conclusions. If someone lives near a good university, they can often access the material. Second, the detail provided by the literature may be more than what some questioners seek. Perhaps they want just a general overview. Here, I take the philosophical approach of StackExchange in that answers are provided by experts. If the answer I provide is not satisfactory for the questioner, then she or he can choose not to accept my answer. I take the long view of StackExchange as a resource for many users, not just the asker of a particular question.
That said, I realize that many others are quite content to include references to a wide variety of sources. I certainly do not want to discourage others from answering with whatever sources they wish to include, nor would I want to violate the spirit of StackExchange. If, however, I feel that I can contribute an answer that expands on or offers another perspective to previous answers, then I'll add my thoughts.