The possibility of error exists about the information in/content of all historical records. It is at least in part the reason to conduct an exhaustive search, hoping to locate information that is in conflict. Once a conflict has been identified and documented, I conduct research and apply logic and reasoning to resolve the conflict and assess its importance.
You are inquiring specifically about the content of [published] obituaries that you transcribe into your family file.
- First things first. Anytime you quote someone else, consider your scholarly and legal obligation to accurately represent and attribute their work. So, I start there--as much as possible, I prefer to quote material directly, warts and all.
- Faithful transcription in hand, work to understand the information it provides in the context of your common knowledge and other sources. Common knowledge concerns things like mis-spelled words that result from an editors error or oversight. Often even those misspellings add a little character to the materials and don't need to be altered. The spelling of proper names, though, is less likely to be a misspelling that wasn't caught by the editor--it might be (a) another variant (correct or not) of the same name (Kittel vs Keitel) or (b) a changed name/different or confused person (Morse vs Morris). In a historical context, my logic would rely on other sources of information to determine whether it was a name variant or a changed name/confused or different person. (This is a process of resolving conflicts.)
- Given your understanding (and documentation), then decide how to best present that information to others. When quoting material, you have the option of using editorial brackets to add/change or notate items within the quote. Many people use "[sic]" to call attention to what might be an error. See Wikipedia, Sic. This notation, sic, is Latin and has the meaning, "thus it was written." (For those of us who don't read Latin, it can also mean "Don't blame me, I didn't do it.")
These various points are easier to answer in context.
The Genealogy.SE question about Marie (Horjas) Kuchar is a real life example of an obituary for Mildred Weber which raises spelling and name/identity issues.
- As to Maria's full maiden name, we were not able to locate earlier "official" documents such as her birth and marriage record. About her surname, we were able to locate what seem the death records of her mother and father. The man who seemed to be Maria/e's brother was the informant on the father's death certificate--both their surnames had been spelled "Horejs." As to her given name, we found sources that referred to her as both Marie and Maria.
- The obituary for Mildred (Kuchar) Weber referred to "her ... sister, Laverne Weber." When we inquired if two sisters had married Weber men, Luke wrote, "I could ask the living siblings the next time I see them. I think it is an error, but I don't really know."
- While we were not asked, the obituary makes reference to "19 grandchildren, 40 great-grandchildren; and two great-great grandchildren." Errors in such tallies have surely been made over the years, and, as suggested by your example, editor error and oversight have surely occurred.
As with the Weber obituary, information is subject to error and the certainty with which we assess the information is not nearly so black and white as we would like it to be.
Using the Weber obit:
- In my family file profile of Mildred (Kuchar) Weber, I would quote her obituary as it was presented, warts and all.
- In the profile about Mildred's sister, Laverne, if I didn't know that she had been married to a Weber, then I would notice the conflict by creating an entry/tag/pfact in which I would describe the problem; I would source the obituary.
- In the profile about Marie (Horjas) Kuchar, I would create alternate spellings for her surname; to the spelling Horjas, I would source the reference in Mildred's obituary. It's probably a coin toss whether I would create alternate spellings for "Marie" vs "Maria."
One final thought. Having learned the hard way, I prefer not to "editorialize" within the body of quoted materials. I think it detracts from the readability of the entry. If I felt that was necessary, I would probably opt not to create a block quote for the obit. Instead, to write about it and incorporate select quoted phrases from the obituary.
P.S. For family historians, there are probably more issues to the notion of "perpetuating errors." This is in part because some of the information we record and share we have yet to prove (indeed, some we may never prove). I'm working on a question along those lines.