Friday, May 1, 2009

Correcting Records on

Many subscribers complain about the indexing errors that occur in the databases, but how many of us do something about it? Fortunately, permits users to add a correction when a name is not transcribed accurately, is hard to read, has a maiden name, has a name change, etc.

I did this today when I could not find Ranslow Smith (1805-1875) in any US census from 1850 to 1870. The reason was simple - he is "Rauslow" in 1850 in Dodge County WI, he is "Rauslon" in 1860 in Dodge County WI, and "Ronsle" in 1870 in Taylor County, IA. If he had lived longer, I'm sure the name would be wrong in the 1880 and 1900 census records too!

Here is what I did today for the 1860 US Census entry. Here is the search result for Ranslow Smith:

In the left column is a link for "Comments and Corrections." I clicked on it:

The page above permits the user to correct Misspelled or alternate Names, Add a Comment, or Report an Image Problem. I clicked on "Misspelled or Alternate Names:"

On the page above, I added the correct Given Name and Surname, then clicked on "Transcription Error" from the dropdown menu of choices (the other choices were Incorrect Original, Nickname, Birth/Maiden Name, Name Change and Variation).

I wrote a short note in the Description box explaining why I think the name should be Ranslow and not Rauslon:

Finished, I clicked on the orange "Submit Correction" button and received this message:

It says that the correction will be added to the Search Index in a few weeks, and the little yellow triangle indicator thingamajig should show up on the result page shortly. Note that this will add the name "Ranslow Smith" to the Search Index and will not replace the "Rauslon Smith" name in the Search Index.

I did all of this in the "Old Search" because of the "New Search" problem I highlighted yesterday in Quirk - Different Results for Old Search and New Search.

I noticed one more error on the Search Results page. It says I searched for "Smith born in New York in 1803 and died in Wisconsin." I did not put a death place in the Search box. I put "USA" and "Wisconsin" in the Residence box for the 1860 census. This process error needs to be fixed too. I don't know if the Search is affected by it - probably not from my experience.


Anne Mitchell said...

Randy, that was an excellent explanation. I think you included everything.

I'd love to hear from you and your readers on what else they'd like to see in the correction/augmentation process.

Geolover said...

Randy, the corrections of transcript/index entries will appear marked by a little blue tree icon, not by the yellow triangle.

Ms. Mitchell's blog readers have been pointing out over the past 9 months that there are many other items incorrectly rendered in indexes.

Just as a few examples:

In the 1850 US Federal Census enumerations, the abbreviation "Ia" is nearly always for Indiana, not Iowa, as State of birth.

US Federal Census enumerations contain many erroneous place-names as the enumeration locations. Some are inexplicable entries for nonexistent places, some just minor but significant typographical errors that can affect a user's search results.

The UK Birth-Marriage-Death place entries as indexed are inexplicably different from those actually given in the records. A similar approach affects searches for 19th-century UK Census enumerations.

A wide swath of indexed entries for the recently added dB "Delaware Marriages 1744-1912" give marriage dates 100 years earlier than the date of the record, and correspondingly derive erroneous purported years of birth for many of the parties.

Similar errors occur in Tennessee Marriages and other vital-records dBs.

Eileen said...

Thank you, Randy. I never knew there was a way to request corrections errors we thinkwe find during our research. Your explanation of how htis is accomplished was very clear and easy to follow. I will be sure to use this information if I uncover an error.

Julie said...

I too am frustrated that people don't take just a few moments to send in a correction and essentially help other researchers. I posted about this back in February: