When my ancestor died in Illinois 1903, she was survived by her husband who died in 1904. His death record, probate record, tombstone, and obituary all make it clear he died in 1904 as all her records make clear she died in 1903. Their years of death are correct--there's nothing wrong with them or any "trick" about when they died.
The ancestor's husband is not mentioned in her will (which is not unheard of) and, more importantly, he is not listed in the court record listing all of her heirs-at-law at her decease. In every other record I've seen during this time in this state, the surviving spouse is included in the court order establishing heirship as the spouse is an heir. His failure to be listed as an heir is "negative evidence" that he was not her husband when she died in 1903. His absence from the record is what makes it "negative evidence."
It's negative evidence because we are using the fact that something we expect to be a part of the document is "not there" (ie. his name is not on the heir list). It is his absence that is the evidence. The difficulty for some is that there's nothing which can be quoted as a part of a citation because his name is not on the list.
Our commentary regarding the order finding heirship should indicate that the husband's name is not on the order and then cite the order (including name of the estate being settled, date of record, packet number, court, etc.). We can cite a record even if there's "nothing" we are quoting from it.