I am a historian of modern Central and Eastern Europe who works on cultural and social themes, so my answer is obviously shaped by my field. The question you pose has two different answers. On the more theoretical level, we can dive pretty fast into a sea of nihilism: Can any source really be collaborated? How can you know that an event remembered by many or reported in real-time was not just one rumor or fabrication that quickly became widespread? Can you ever trust someone to write his diary truthfully?

In practice, historians only rarely can afford to dive into the abyss. We do our best to verify the truthfulness of what we write, and the efforts spent in doing so are larger when the event is of greater importance. My current book project, for example, studies the transition of European youth to adulthood between 1890 and 1968. Many of the events I work with are small, like one wedding of an ordinary couple in Budapest, and the greater picture is of more importance than any particular example. My writing is guided by the assumption that the truth emerges, probabilistically, by the use of many such events and not just one.

My colleagues who study political and diplomatic history will certainly differ. In their case, what was actually said in a certain meeting can be sometimes important to the level of the comma. The same goes for battlefield movements for my colleagues in traditional military history. When the particular event holds such high stakes for the argument, the writing will usually reflect the amount of certainty the author has in what they are saying as well as include the necessary qualifications.

If you decide that living with this uncertainty is too much for you, one solution is to study representations. That means not talking about what happened but about how people wrote about what they thought happened. This means that you won't be writing about sexual behavior in 19th-century Switzerland but on how the newspaper, literature, and the like talked about sexual behavior. This is a much more sound empirical path to take but also, arguably, a less exciting one. Historians usually balance between studying representations and striving towards understanding the historical reality itself.

Bottom line: Yes, we do. Good writing reflects our certainty or uncertainty, but when books depend on hundreds of micro-events like the death of an anonymous person in a random Transylvanian village, the relative certainty might come from the sheer numbers and not from detailed investigations in writing of each particular events. (It does happen too, sometimes, in the genre of microhistories that is popular in the study of early modern history, but that's a subject for a completely different answer...).