His post stems from a discussion on a professional genealogists’ mailing list. A list member experienced with NARA records did a spot check: She noted the first 25 names on a NARA microfilm reel of Civil War pension index cards and searched for those names in Ancestry.com’s pension index database. She found just one of the names. (I can hear you thinking “I knew it!”)
The researcher said the cards that didn’t scan well from the microfilm were left out of the database (Ancestry.com’s source information states 1 percent of the cards are “missing;” she puts the percentage higher).
The researcher also questioned the wisdom of scanning colored documents in black and white, pointing to Footnote’s Civil War widows’ pensions project.
A NARA staff member explained that partner digitization projects use original records or the highest-quality “master” microfilm and are subject to quality controls. Other, non-partner projects may have digitized records from second- or third-generation film, resulting in poorer images.
He also said NARA does make original records available, even after they’re digitized, to “researchers who need to see them.”
A respondent from Ancestry.com commented that the microfilmed Civil War pension index cards were particularly difficult to scan because some cards were on dark paper, and the technology available at the time was inferior to today’s.
See Seaver’s entire post here. He raises good questions at the end.
It’s easy and comforting to assume genealogy databases have every surviving document in a particular record set. This is a reminder that’s not always the case.