When it comes to census mistakes, a lot of them are actually the fault of the transcriber. And the best way to look for them is to compare the questionable spelling with the spelling of more clear words to see how the writer formed his letters. There's a "Daniel" Evans in the 1790 census for Washington County, Pennsylvania, which when you look at the spelling of other words you can see it's really a David Evans. I figured it out only by looking at the neighbors on the page which matched those of my known David Evans. I'd assumed it was a mistake by the census taker until I looked it again and that was just the way the census taker made his letters, so it was a transcription error although one I don't really blame the transcriber for since might have made the same mistake.