Sure, why not? It depends on what you are using them for. As you note, they aren't to just be taken at face value ("and thus, we see, that aliens did invade Earth in 1897"), but they can be used for many other types of claims ("and thus, we see, that the idea of atomic bombs well pre-dated the discovery of nuclear fission"). Fictional works can be primary sources (most academic works are secondary sources), indicative of ideas and cultures and sometimes even practices of the past. They have to be used with just as much care as any primary source, and with an awareness that their deliberately fictional nature does constrain the kinds of claims they can be used to support. So you cannot use, say, Elizabeth Gaskell's 1853 novel Cranford to talk about the actual effects of the railroad on British small towns in the 19th century, but you can use it to talk about how the railroad was perceived and discussed by people living at the time. One of course has to take care in such work; one shouldn't use Dickens to talk about how people thought about England at the time of Dickens' writing, since most of his writing is set in a somewhat earlier period (thus it is how Victorians thought about life in Georgian England; somewhat similar to imagining how you would look at a novelist today who set his or her work in the early 20th century, it would tell you less about the early 20th century and more about 21st century perceptions of the early 20th century, which are not the same thing).