Thursday, November 18, 2010

Answers to Indexing Illustration

These are the eight presidents from yesterday’s indexing illustration:

Presidents Answer

How’d you do before knowing something about the names?

How'd you do once you knew they were presidents?

The point is that context makes it easier to read names.

For us less experienced genealogists, the context is usually information known for several family members: names, genders, probable birth states, and estimated birth years. Once you have matched lots of known information with what is written legibly, you start to approach the illegible names with some confidence. You match expected letter forms against what you see. Pretty much without thinking about the complex brain gyrations, you "read" names that cold indexers can not.

For more experienced genealogists, you build up a name probability dictionary in your head for particular times and places. You pull from it to compare names against letter forms, allowing you to "read" names that others can not.

I conclude that cold indexers will never outperform someone with more context.

What do you think?

12 comments:

  1. Your 'knowing the context' illustration with the presidents' names proves what I have always believed - each county should be indexed by people whose family has lived there for several generations. They will recognize the names in spite of the handwriting.

    ReplyDelete
  2. Yeah, context is very helpful. I love how many people in the census are from India in 1850-1880. When census enumerators wrote "Ind." and the person is named Charlotte Smith or something like that, and yet the birthplace is transcribed as India, I just have to laugh. I wonder where Ancestry's outsourcing the transcription to?

    ReplyDelete
  3. Indexing is even more difficult when the original is spelled incorrectly. Presumably your third example was intended to illustrate this? :-)

    ReplyDelete
  4. This is so true. I have done indexing and it is sometimes hard to read some of the names. I make corrections on Ancestry's index for my families that have been mis-indexed, but I can clearly see how the indexer came up with their view of the name.

    ReplyDelete
  5. I've become somewhat obsessed with correcting my ancestors' indexing errors on Ancestry.com. Some of the French Canadian records I search were so far off that even quite experienced local genealogists could not find the records! It makes me feel good to know I'm helping the next generation of researchers looking for a common ancestor.

    My next question is: What to do about records that are not indexed at all?? I found one essential birth record by sheer time & persistance, because for some reason there was no indexing.

    ReplyDelete
  6. I have to admit to looking at other sources for that "context" when I can't read something in a record. For example, if I'm indexing 1898 Illinois Marriages and can't figure out a name, I look for the couple in the 1900 census. I'm not sure if that's kosher, but it sure helps me transcribe unusual names. I also select indexing projects from locations where I have some research experience. It really does help.

    ReplyDelete
  7. Yes, but if we had to wait for someone in "each county" "whose family has lived there for several generations" to create indexes for the genealogical records we need, we would be waiting for decades or longer. I would rather have millions of records available sooner with a marginally lower quality index than have to wait for the perfection that so many seem to be seeking.

    A good search engine (like the one on Ancestry.com) allows you to tweak searches looking, for example, for all males born between 1870-1880 in Ohio, living in Multnomah County, Oregon - without having to enter a name at all. I can browse through four or five pages of search results and find my people amost every time. In addition, light indexing allows images to be posted in smaller chunks of data that I can browse when I find other family members in the proximity - which I would hope most people do anyway.

    Lastly, I'm convinced that if all of our research really was as simple as the push of a search button to find perfectly indexed records, many of us would get bored very quickly. The challenge of it all is part of what keeps me engaged. And, I feel more connected to those I have had to struggle a bit to discover.

    ReplyDelete
  8. It's absolutely true that cold indexing is very difficult and error-prone, having done a lot of that myself (undoubtedly with lots of errors...). But I'm really glad that there is so much available out there, errors and all. Just means you have to be very smart, flexible and creative, when searching. And means you learn long lists of "likely errors" for different kinds of names, times and records. At least the info is *there* when you finally figure out where *there* is.

    ReplyDelete
  9. And don't forget that spelling counts, too. His name was Ronald Wilson REAGAN.

    Happy Dae·
    http://ShoeStringGenealogy.com

    ReplyDelete
  10. whoa! you mean I'm NOT from India? so much to change in the file now! (hee hee) Ancestry does allow you to suggest changes for a record where there is a transcription error, so that helps a little bit.

    ReplyDelete
  11. You have misspelled the name of Ronald Reagan.

    ReplyDelete
  12. Really interesting exercise. I printed your column to take the test in another room, and accidentally discovered -- if I hold the page almost flat at eye level, with bottom right corner pointed toward my nose, I could READ all the names except *3, and that one I guessed. (Yes, knowing they were presidents,helped.) No, I'm not suggesting indexers do something similar. AI, try this for fun. Your point, about indexing, is well taken, however. Cheers, Dolly in Maryland

    ReplyDelete