Revealing issues with transcriptions
As I work down through family lines, I see many transcription errors issues. For example, the 1850 U.S. Census for Marion County Ohio seems riddled with them, especially in state abbreviations. I don't know if these are man-made, or computer interpreted. One of the issues is that someone turned UP the contrast setting which has made these images difficult to read.
In any event, the tool allows us to go in and find the incorrect transcription and at least put in a correction. (But Md is Maryland, and never Missouri)
Answers
-
Many - if not most - of those errors were introduced by the placename algorithm that picks a random place that begins with the same letter. We truly cannot rely on the placename in the index for any record set.
Just yesterday, I was working in the 1840 census in Georgia, USA. My ancestor lived in Walton County. His census record has been "moved" to Walker County, Georgia because that county comes first in the list that starts with the same letters.
Even worse, in my search, I saw a similarly named man who lived in Liberty County, Georgia. He's been transplanted to Philadelphia, Pennsylvania. I can only guess that "Liberty" and "Philadelphia" are considered to be the same by the algorithm.1 -
Today I encountered a consistency error for Upper Arlington, Franklin, Ohio. The couple moved to UA in 1921 while it was a village (in Ohio, a village is an incorporated community with fewer than 5,000 people). However, the consistency error claimed that UA only existed after 1941. 1941 was the date that it was incorporated as a city (population greater than 5,000). So under the standardized name rule, there were 21 years where this place didn't exist? I was able to send the correction through to get it listed as a village from 1920-1941. To me, this profile grading is a win.
0
