Talk:FamilySearch Indexing: US—1940 Federal Census, Project Updates

The talk pages are not supervised by Indexing support -

What do you do when he name of the state is written in the city field? (Please see the Wiki Project Updates page for the answer under the Place of Birth and Residence fields section)

Hawaii Census - Race
The census form shows Caucasian - Cau. The indexer put white. How should this be arbitrated, Caucasian or White?

On the Drop Down Menu Cau/Caucasian is not an option, so I would go with White. That is how I would Index and Arbitrate it.

What to do with crossed out records?
Just got this batch back from Arbitration. Basic Indexing Guidelines state that "if information is crossed out, not replaced, but the original information can be read, type the crossed-out information", which I did, since there's no project specific instructions that contradict this guideline. The arbitrator went through and marked lines 1-29 as blank, which is more than half of the work I did. I'm willing to accept that I did this wrong but unless I'm completely misinterpreting the guidelines I'm pretty sure I'm in the right... anyone able to clarify? This probably isn't the only batch with crossed out information like this.

Comments with feedback when reviewing arbitration?
I had a very difficult when indexing a page from Wisconsin. It was loaded with Norwegian last names, poor handwriting and bad focus. My solution: I found the same street in the 1930 and 1920 census where the image quality was better and found the same families and how they were spelled back then. However, the arbitrator did not always choose my entries, even though I was very confident of the ones that were hard to read because of my research. Is there any way to comment to an arbitrator when sending feedback to explain why your choices were made?

.......

I am an Indexer and an Arbitrator. I agree with everyone's observations. I just want to comment that the previous poster is good to have mentioned looking in earlier censuses to compare name spelling. Remember if the name is spelled wrong on the Census we have to copy it, but using earlier Censuses to look up names for coroboration of what you think is written is very helpul. I do it at some point on almost every page when Aribitrating. The Lookup Lists are also very helpful for getting coroboration of what you think you see in each field. I also use the Internet to look up correct spelling of cities and counties I am not familiar with. Cannot correct the county if the wrong county was entered, but can correct spelling on all the geographical names.

........

I have the same concern. I'm an arbitrator but do indexing sometimes. I sometimes get low scores on my indexing because the arbitrator was mistaken. This is very frustrating and discouraging. I'm not going to quit, but I find I'm doing less and less. Again, it is discouraging to be doing your upmost to be accurate and have it undone by someone else.

.......

I agree, I go through the same pains trying to get it right. There is a Feeback option, however, I am not sure it actually does anything. You can only check a box that asks someone to review the arbitrator's changes, you cannot provide any actual information as to why you asked for this review. Some arbitrator's are good, others should be pulled from the program.

........

I entered data for a Texas county re property taxes and the aribtrator completely ignored the instructions and replace the owners of the land with the original grantees of those pieces of land, rearranged the order of the entries to completely ignore the order on the actual pages, and generally screwed it up. The result was a 34% correct for the 2 pages done after this alteration from reality. Although I've entered a check mark for the feedback choice for well over half of these fouled up entry changes, I doubt that the feedback is more than spot checked. The resources simply are missing for this. There needs to be a way to request that a SECOND arbitration take place in cases where the percentage is unreasonably low, since clearly someone who gets in the upper 90% range all the time will not get a 30% rating on a sheet unless someone is misreading the instructions. In this case it was the arbitrator, and there is NO way I can get anyone to really look at what was done and explain to whoever did the arbitration what was incorrect with the way he or she did this work.

.......

When I signed up to be an Arbitator, I read somewhere in the process that the Feedback is looked at, and that if an Arbitrator is consistently having problems, then they will be asked not to continue. I hope this doesn't happen to me, but I think it is a very good policy, and hope that it is true. I think it is very important to continue to check the feed back box when you feel that instructions are being disregarded by the Arbitrator. When checking my own arbitration results after Indexing, which I also do, I tend to give it to the Arbitrator on the spelling/number issues, when it is close, as that is often opinion, and the Arbitrators are choosing between 2 Indexers opinions. When Arbitrating, and on the occasional times that it appears both Indexers are obviously off, I definitely check other sources, like earlier censuses, before entering my own opinion. At some point I hope they configure the system so that Arbitrators can see their feedback. If I have missed how to do that I hope someone lets me know.

.......

This page indicates that it is not 'supervised by Indexing support'; it obviously is not even visited 'by Indexing support' - there appear to be unanswered questions and absolutely no discussion from anyone sounding like they are from 'Indexing support'. It appears that we are all talking to the wind.

If there is any 'official' monitoring these discussions, please reconsider letting just anyone be an arbitrator. It is a real shame to have the work of 'good indexers' destroyed by 'bad arbitrators'. You will end up losing at least some of your 'good indexers' if they continously see their efforts ignored and overwritten by bad data. Not to mention that family search is not helping the researcher if your indexing is not as accurate as possible.

..............

What about questions... not just comments. In one of the records I indexed, the census-taker had put all of the family numbers in the wrong box. The arbitrator went through and changed all my family numbers to blanks. I'd like to ask if I should have done what I did. It seemed so obvious and I'm thinking it would be hard for someone to seperate the entries into families if they were all blank. Let me know... it result in my lowest arbitration score ever!

............

I agree with all of the above, a comment box would be nice when clicking the feedback form. I've had several pages come back with "corrections" that should not have been made: "Wm" being changed to "William" even though the instructions say to keep abbreviated names abbreviated; line numbers being entered for blank lines when instructions explicitly state otherwise; cities/counties/states/places of birth blanked out when there are ditto lines; and those are just off the top of my head. A vast majority of the "mistakes" in my indexing have actually been correct in the first place and are the result of an arbitrator not reading or following project instructions or updates. Another suggestion would be to allow us to review our arbitrated records (as we already can), and if the arbitrator clearly made mistakes, clicking "please review" should immediately send the record BACK to the arbitration queue, along with our comments in a popup box. If FS is worried about comments being "heated" or inappropriate they could add some ticky boxes to the Feedback popup with common issues. If anything, this would allow FS to see which arbitrators are making a lot of mistakes and perhaps bump them back to only being able to index if they dip below a certain percentage of agreement. Yes it would take slightly more time to get records out there but I don't think quality should be put on the backburner in favor of quantity!

Not only that, but I'm sure the arbitrators would be just as upset as us when they read the results and then take the time to READ the rules!

"Ab" after the given name
I understand that "Ab" means the person was temporarily absent from the household on April 1, but should it be indexed in the Titles or Terms field? I'm guessing it shouldn't. I think I remember not indexing it for the 1930 census, but I just thought I would check. Thanks!

The directions tell us NOT to index the AB.

Index the "R" or type out "RURAL"
When the City of Residence has an "R", should the "R" be typed or "Rural"? I've been arbitrated both ways. What is correct?

I read in the directions that it should be typed as it is written. Is the arbitrator changing your answer to match the form?

Alleghany County, MD vs. Allegheny County, PA
I have been mis-arbitrated several times on this one.. I was pretty sure I was right, since I grew up in Western PA, but I researched it to verify...

Although they both have the same long a pronunciation, the correct spelling in Pennsylvania is Allegh e

It ends with - a ny in Maryland.

Allegany, Maryland is the correct spelling.....Allegheny, PA is correct for PA

"Same place" notations in New York City [- Agree (does anyone from Family Search actually read this discussion board?)]
Project instructions state that when "Same place" is noted in Column 17 (City of residence in 1935), columns 18 and 19 (county and state of residence in 1935) should always be marked blank. This instruction makes sense in most places. New York City, however, encompasses 5 counties within the city limits. Therefore, someone may live in Brooklyn in New York CIty (Kings County) in 1940, but have moved there from Manhattan in New York City (New York County) in 1935. So they are in the "same place" i.e., the same city, but in a different county.

Indeed, in the pages I have been indexing from the Bronx in New York City (Bronx County) I have found numerous incidences where people have moved from one county within New York City to another. I have therefore -- on New York City census pages only -- when "same place" is noted in column 17, been filling in column 18 whenever the enumerator has done so. Column 19 (State) is always the same, so I mark it blank. With "same house" notations, I mark both 18 and 19 blank.

I expect that when arbitrators get to these pages, they are going to "correct" them according to project rules, but I don't know how to prevent this. For researchers, knowing which county someone lived in in 1935 can be valuable information. It's a shame to have it lost to those who don't look at the original images. [Blanking Column 18 also misdirects research when trying to back-track where their ancestor's lived during the State's 1935 Census, and possibly the 1930 Federal Census]

Just a comment. Please remember that the geanealogical researcher should just be using our indexing/arbitration only as an index to locate the name. To record the data they should be going to the original source, the Census page. So if the county information is on the original Census page, the researcher will see it. It adds to Arbitration time to have to go back and "Blank" any county data that was left in a Same Place record. Please follow all guidelines and updates.

== Thanks for trying to help, but unless the field instructions say so, please type what you see ...and if you are an arbitrator, please do not change our blanks (as the directions unfortunately tell us to do when someone lives in the same place) to a field value that was written for county and state! ==

I am relatively new to arbitration, but in just the last few days I am totally surprised by the number of indexers who seemingly try to enhance what is written on the census page. This includes:

- renumbering lines after encountering a blank line (so if line 17 is blank, then the indexer places 17 on the record for line 18, etc.) Please just omit the line number of the blank line and use the line number from the form for the next line. Also, if the form (all the "B" forms) start with line number 41, you need to use line number 41 as well. We need to stay consistent with the line numbers on the image so researchers can find the correct line on the image quickly and accurately.

- placing zeros in front of single digit line numbers ( 01, 02, 03, etc.) - please don't.

- pulling down the city, county, state from the header into blank fields even though the instructions say not to do this

- correcting the spelling of names or writing out abbreviations to names when the instructions of certain fields do not say to do this. Some do, some don't. If this is hard to remember, you might want to make yourself some notes about when you should. The field after the given name is for identifiers, like Jr, Sr, Mr, Mrs, etc. The instructions clearly state to tab over this field when it is not needed.

- not looking at the previous image when needed to fill in dittos (lines, etc.) at the beginning of your image. We need complete information, even if that is only found on the previous image.

- placing the surname in the given name field and vice versa.

- TYPING IN ALL CAPS - this is a no-no. It looks bad when ALL CAPS and Mixed case are placed in the same record - especially as the result of the arbitrator choosing some of each to get the best info.

- use Control-B or the proper button to mark fields blank. Simply typing the word, "blank" is not the same as the system marking the field &lt;Blank&gt;

We all want access to these indexes, but please take the time to actually read and become familiar with the project and field instructions. Indexing badly slows down the process of arbitration and in turn slows down the availability of the index. We are all human, and we all make mistakes (I am still learning about this project), but let's all do the best we can!

Thanks.

Remember that the "rules" for how something is entered is not always determined by FamilySearch. These records we're indexing come from partner organizations such as governmental entities. THEY most often are the people who determing how problems should be handled and what should be entered, how it should be entered, and even what should be entered or indexed.