Indication for work submitted.
It would be extremely helpful for all indexers to know how their submitted work is rated.
E.g I have just indexed a batch of burials for Middlesex, England for the year 1593, which someone has indexed as christenings and the
names entered were sheer guess work. I am sure that all of us would appreciate a % rating for each batch we submit. That would give us
an idea of how well we are doing, otherwise how are we to know ? Please give this some thought - to me it makes complete sense. Regards Jean.
Comments
-
Jean,
What you suggest would be a wonderful warm fuzzy for new indexers or even old experienced indexers who are starting to work on a new project. In the old application based Indexing process providing that kind of feed back was a feature. However, in the new Web Indexing process that feature was dropped. I have heard that one of the reasons it was dropped was that it created some contention between indexers and what were then called arbitrators when indexers did not agree with the arbitrators conflicting interpretations.
As I understand the Web Indexing process there is a way for FamilySearch to counsel indexers who consistently have their work corrected by a reviewer, but it is only used in the most serious of instances.
0 -
Since we do not have that anymore might I suggest that you make use of the ward temple and family history consultants to help you see where you might improve. One thing that was helpful to me is when we had people index together and help each other. We may not be able to do that in person right now but by using the shared batch code you can work with each other at your own computers at home. Some wards do set up zoom conferences for indexing groups to index together. You may be able to come up with other ideas that fit your situation better and be fun as well.
1 -
If your ward Temple and Family History consultants are well versed in Indexing that is a good suggestion. If they are not or are only vaguely familiar with it you might try contacting the Stake Indexing Director, that is assuming that your Stake has one.
0 -
The percentage rating that we used to get and seeing the corrections was a double-edged sword. Oftentimes the arbitrators did not look at the instructions or the field helps and made more errors than the indexers made. This caused a lot of confusion. I fear the same would be true today, especially when nothing more than indexing 1000 names allows a person to become a reviewer. At least with arbitration, it was hoped that people got training before receiving their arbitration rights from an indexing director. It was quite irritating to receive a low score from a person who had no clue what they were doing. I know many who quit indexing because of this.
Not everyone who indexes is a member of the Church. For many of us, talking to a Church consultant is not feasible. But, if you have a Family History Center near you, that is definitely an option since you can visit without being a member. I also suggest private messaging a person you trust, or reaching out to someone in your indexing group to work with you. Usually two sets of eyes on the project helps a lot!
0 -
My questions are more to deal with knowing how the indexing process works than how my indexing is rated.
For example, I try to be thorough in indexing and would like to know if my index ends up in the final index that is submitted for review. How many indexing passes are done? Does any of my indexing survive to the final index submitted for review? How can it be determined whether my guess at deciphering a field is any better or worse than another?
Obviously the answer could depend on educating myself on the handwriting/type of document I am indexing but even then a brand new indexer may see something more correctly than I did or not. The same goes with the review process. That's why I would like to know the process to understand how FamilySearch gets the best possible results for researchers to find the document as easily as possible.
Yes I am more of a researcher than indexer - but I think that's because I view the researcher as the audience I am indexing for. I was told today on a question that I should take off my researcher hat when indexing - why? If I leave the hat on shouldn't that make my index better?
That's mainly why I had confusion about why I cannot enter multiple guesses at deciphering a record field. If I can see discrete possibilities for a name why not enter them on the first pass of the index process, have another indexer pass see what they think , etc - or at least make differing indexing options available for a reviewer to review and make a decision - and even then how to make a final decision?
I do agree with @Melissa S Himes - reviewers should be at a significantly higher level of deciphering accuracy than the indexer otherwise who is to know the final index is worth something?
For these reasons, again, I would just like to know more about the established process - maybe this is already taken care of behind the curtain - but otherwise it leaves the record more difficult for the researcher to find - which is the frustrating part I see. It does seem to make finding the record and editing variants on the document image by the researcher more important - but it's more difficult to try to be an Indexer by Edit on the published document for those that are editable. I'm trying to just do my best as an Indexer and hope that's good enough.
1 -
The process is as follows:
An indexer enters the information that is found in the image for each record that the image contains and submits the batch. A reviewer checks what was entered against what they find in the image and makes corrections if necessary. Those corrections can be changes in the spelling of words or they can be corrections that were necessary because the indexer did not follow the instructions for the project, things like indexing information that the instructions said should not be indexed or not indexing information that the instructions said to index. If the reviewer changes fewer than 20% of the fields in the batch the batch is considered correctly indexed. If 20% or more of the fields in the batch were changed a second reviewer will review the batch. If the percentage of changes is less than 20% of the fields the batch is considered correctly indexed. If the percentage of changes is 20% or greater a third reviewer will review the batch. If the third reviewer changes less than 20% of the fields in the batch the batch is considered correctly indexed. If the percentage of changes is 20% or greater the batch is sent to the FamilySearch staff for a final review. The FamilySearch staff can see the entire history of the batch and makes the determination as to what will be considered the correct interpretation of the information in the image.
Regarding the entering of multiple guesses as to what the information in the image is, the process does not provide a way to do that. I realize that as a researcher it seems like it would be a good thing but there is no mechanism for doing so. I have seen people put Or between two guesses as to what a word in the image might be, but that notation is reserved for entering aliases, name changes and cases where the image actually contains both words. The only mechanism that is available to note that a character or characters cannot be deciphered or might be one of two or more characters is to use the ? symbol for a single character and the * character for multiple consecutive characters. Using these symbols provides the computer with a better chance of matching the word to what a researcher enters in a search because it means that they can match any character in those positions.
0 -
@J C Bingham Good information to know. From that I think the process sounds pretty good - however - that 20% seems pretty important. Depending on how many records are in the batch that could mean a couple to a dozen or more incorrectly indexed records that slip through the crack - relying especially on that first Reviewer pass.
0
