Let me throw in another wrinkle:
Online reports of the gravestones for Peter and Elizabeth Baker Wright, in Shaftsbury, VT, indicate a death date of 1821 for Peter and 1819 for Elizabeth, with their ages given specific to months and days, as was common at that time.
It seems to me that there is a mix-up here, as Peter could not have died in 1821 as he was not recorded in the 1820 Census - his widow Elizabeth was. The enumeration date for the 1820 Census was 7 August, 1820. If the Bennington County census taker recorded his data correctly, that would mean that Peter Wright died before that date, and Elizabeth lived past that date. Even if the census taker recorded who was in residence on the actual date he visited the Wright home, his return would indicate Elizabeth outlived her husband.
Taken even further: if the name and specific age of each Wright is correct on the gravestone, but the death dates have been reversed between the spouses, that would mean that Peter Wright, aged 80 - 1 - 2, who (may have) died on 16 August, 1819, would have been born on 14 July, 1739.... which puts his birth well within the reported pregnancy of Temperance Seaman Wright, second wife of Joseph Wright.
I know, so many 'ifs', but the census data for 1820 would seem to indicate that the death dates given on the Wright graves are not correct.
I'd welcome more analysis - or better yet, email@example.com