Tuesday, April 17, 2012

FamilySearch Indexing Challenges - Post 2

I've been reviewing my FamilySearch Indexing successes and, um, challenges.  The good news is that I finally achieved 100% on a 1940 U.S. census batch - the one for Beverly Hills, California where the handwriting was very clear and the enumerator actually knew the place names.

Then I checked the previous batch that I indexed for Antelope township in Los Angeles County, and had these noted errors:

*  1935 Residence was enumerated as Trinidad, Los Amis, Colorado.  I looked up where Trinidad was, and it said Las Animas county in Colorado.  The arbitrator selected "Los A"  Not "Los Amis"  Oh well...two dings.

*  Surname looked like Pelater to me (quite blurry), the arbitrator selected Pelster.  One ding, perhaps correct.

*  Surname looked like O Dane to me, arbitrator selected ODane.  Two dings, perhaps correct.

*  1935 Residence was enumerated as Alamonte, Los Angeles, California.  I looked up "Alamonte" and there was none in Los Angeles County, so I indexed it as El Monte, which was in Los Angeles County.  The arbitrator selected Alamonte.  Two dings.

*  1935 Residence looked like R, Caddeu, Oklahoma.  I looked up Oklahoma counties and the only one starting with Cadd was Caddo, which is how I indexed it.  The arbitrator selected "unreadable."  Two dings.

An earlier batch had these errors:

*  Color or Race - the name was Nielsen...the field said "Indian."  I selected American Indian from the list, and the arbitrator selected Indian.  Oh well.

*  Line number - the line was blank, so I used Ctrl-Shift-B to mark the entire line blank as previously stated by veteran indexers and the project directions.  The arbitrator selected the line number, which is clearly wrong.  Four dings.

So, out of the 13 dings above, I think I was right on 10 of them.  But I guess I cheated by looking up the actual place names so that others won't have to do it later.

My arbitration results are up a bit - I'm at 97% for the whole time, and have an average of 98.6% for the 1940 Census batches (but it should be higher - right?).

The URL for this post is:  http://www.geneamusings.com/2012/04/familysearch-indexing-challenges-post-2.html

Copyright (c) 2012, Randall J. Seaver


Nancy said...

I think they should have done handwriting tests before hiring the census takers! Some of the handwriting is nearly impossible to decipher, even when comparing handwriting in other letters/words on the page. I suppose having several companies index means we get a wider range of interpretations when we search later.

Lauri said...

I did one 1940 batch and scored 93%, so you are doing much better than I am. I was getting 100% on the WWI draft registrations. I'll give the 1940 another try after going through my arbitration results but alas am a bit gun shy at the moment.