Please review your work

A couple of us have spent the last week correcting errors that Chris found in the first 5000 soldiers in the Project 1 database. We're probably a little more than half finished correcting errors. Some of these problems are inputter errors. Some of them are problems of missing data because when the soldier was input, we didn't have a specific field (such as the quality code on the death screen). Some are just glitches that happened in the screens.

Many of the inputter errors can be eliminated by reviewing your work. Before you mark complete, do the following:

• Look at the Household View and see if everything looks correct. As a matter of fact, you should do this after inputting each household.
• Check for duplicate or missing people.
• Check household numbers.
• Check names.
• Look at Update Relations.
• Look for red Xs on the Grid.

Additionally, you should do the following:

• Take breaks. You are entitled to a paid 10-minute break when you work four hours. You don't have to take it all at once. Get up and walk around. Get your blood circulating, and give your brain a break.
• Do not work late at night, especially since you probably had a full and tiring day.
• Do not work if you are ill. I'm not talking sniffles. Your body knows when it needs rest.

One inputter, just spent an hour disentangling a mess someone else made inputting a soldier. There was at least one duplicate person. There were several missing people. URLs were left out. Several household numbers were wrong. I know who the original inputter was. This person is competent. I don't know what happened with this soldier. It could all have been avoided if she had done her work when she was alert and had reviewed it when she was done.

Please review your work.

Progress report as of January 26

Here is our progress report as of January 26, 2017.

	Project 1	Project 2 - USCT
	n = 8,500	n = 4,500
Total soldiers complete	5597	1767
Soldiers completed during last  week	125	18
Soldiers w/o children (to be removed from sample)	534	311
% of completed soldiers w/o children	9.5	17.6
Soldiers with children complete	5063	1456
% with children complete	59.6	32.4

Don't forget your trees!

This is just a reminder that your completed family trees are due when you complete your assignment. The only reason you should delay downloading your trees when your assignment is complete is that you have soldiers out for checks. If you would like to keep your trees in your account after you've downloaded them, that's fine. Please do not make me wait for weeks then beg you for your trees.

Some of you do an excellent job of getting your trees to me in a timely fashion. Some of you do an excellent job of informing me of soldiers being checked, so I know not to expect those trees. I appreciate this.

If you can't remember the last time you downloaded any completed family trees, you might want to check and see if you have some catching up to do.

Blank names

Sometimes a person will only have one name on the census or obit record, or none at all. Follow these rules to input them:

INPUT THIS	WHEN IT LOOKS LIKE THIS ON THE DOCUMENT
(BLANK), Mary	When only first name shows (and last name isn't dittoed from above) such as on a death certificate.
Jones, (BLANK)	When only the last name is given and there is no first name
Jones, (?)	When it says, Jones, ? on the manuscript.
Jones, (WIFE)	When it says Wife in the wife’s line on the census underneath a husband whose last name is given
Jones, (BABY)	When it says BABY on the census manuscript and last name is written in or dittoed from above
Jones, (DEAD INFANT)	When mil info says DEAD INFANT or you find a death record saying DEAD INFANT [you can mark all decades with black x’s even without death date]
Jones, (CHILD)	When it says CHILD on mil info (include Vet’s last name)
Jones, (NO NAME)	When census manuscript says NO NAME
LEAVE BLANK	When it says unknown, not obtainable, not given or just ?  In other words, no names are given at all.

See previous blog post and comments on Sept 14, 2015 for additional reference.


What you're trying to accomplish is to make the name conform to our standard name format rules using whatever information you have. Anything that is not a last name, first name, or middle initial/name should be in parentheses ( ). If the first name or last name is left blank, let us know by inputting "(BLANK)."

Institution name

"Get thee to a nunnery**..."  Info on Institution Name

You should only input the name of the institution from the manuscript if your family lives in the institution.  Frequently there will be line numbers next to the institution name, such as City Prison lines 22-36 OR Sisters of Charity House 64-74. Here is an example where you can see that the family does not live in the institution with the the nuns, which is actually on the next page: Nunnery

*Hamlet

Soldiers without death dates

For the original sample for vets without death dates,the rule was once we had two NF decades after 1880 to stop looking and assume he was dead.  You could mark the two decades yellow and any remaining decades would be marked not searched. If you get a guy like that, we would like you to mark it as you do the children without death dates. 

Mark all decades as yellow unless you have some evidence indicating the vet is dead.

ODD RELATIONS (we all have a few)

There have been several questions lately on what  number to choose for  the following odd cases which are listed in the relation column on the manuscript:

• partner
• orphan
• foster child
• ward
• companion
• hired man
• live in lover
• nurse
• teacher

In these cases, you can mark them as 5.  Remember, when you look at the description for 5, it says "servant, living with."  The comma between them means for servants OR other people living with family. (OK, you got me. One of those odd cases has never really shown up...yet.)

December checking statistics

December 2016 checking statistics

In December, we checked 42 soldiers under our system. I've reviewed all of the checks, and I've tallied the number of differences. Some of these are errors, and some are judgment calls. Here are the categories and the total number of differences for each category.

GRID Errors

MILIN?/MAR? - 2

Missing HH member - 12 (this number tripled!)

Duplicate people - 0

Wrong person - 1

Other - 13

Inferred Relationships

Incorrect relationships - 8 (this number dropped by half)

Census Errors

Name - 44 (too high!)

Typo/Reading/Wrong - 126 (this number went up)

State Code - 0

Missing/Wrong URL - 3

Missing data - 35

Additional finds - 20

Quality Code - 17

Death Errors

Typo/Reading/Wrong - 24 (this number nearly tripled)

Missing data - 32

Missing/Wrong URL/Source - 17

Quality Code - 3

Additional finds - 9

Tree Errors

Missing/Incorrect information/relationships - 7

The total number of differences for all 42 soldiers is 373. This is 77 more differences than November, when we checked 48 soldiers. Some differences/errors affect the data more than others. If we checked other pensions, we'd probably find similar differences. So, in December you checked fewer soldiers than in the previous month but increased your errors. Let's see if we can reduce these differences in 2017.

Another post regarding punctuation

Quick Reminder. Don't ever use apostrophes and ampersands and hashtags.  After you submit the page that has an apostrophe, it won't save anything you just entered.  If you already have saved info and later add a remark with an apostrophe, for example, it won't save any of the new information.

Please don't use apostrophes ('),  ampersands (&) and hashtags/pound signs (#) anywhere when inputting:  Not in the inputter notes or remarks section (e.g. vet's son lives next door), not with names (like O'Neil), not with cemetery names (St Mary's), not in occupations (baker & merchant or mason's helper), not even when it's written on the manuscript page (Dist #2, Pct 2 & 3), not in a box, not with a fox, not here nor there nor anywhere. Thanks.

Regarding Alternate names for Head of Household

Occasionally the enumerator does not indicate "Head" for the head of household.  We all know that in 1880 the Head is left blank, but we are still to consider it a head of household and duly mark it as a 1 in our inputting screens.

Occasionally an enumerator will list the head of household as Husband, followed by wife and children.  This too can be treated as "Head" and marked as 1 in the screen. [However, if wife is listed first as head followed by husband, then she should the 1, and he should be the 7].  There have even been times when the enumerator consistently indicates the head as Father, followed by mother and children.  This too can be treated as a Head (1) and spouse (7).  This way when you input the child as a 2, the relationship relates back to the head of household.  Here is a LINK to see an example.  

Maiden names on family trees

We've decided that you should list women under their maiden names on the family tree. If you do not know a woman's maiden name, please leave her surname blank. We are just trying to be consistent across people and trees.

When you're checking, and you notice that the inputter did not do this correctly, you may remind her/him about the above guideline. There is no need to take the time to change the family tree, in this instance, while you're checking.

You do not have to worry about not listing the woman's married name on the tree. Because her marriage is recorded on the family tree, Ancestry will know the woman's married name is her surname and will use it to generate hints.

Checking responsibilities

Here are a few notes on your responsibilities as checking partners in our monthly checking system.

• You must send soldiers to be checked to your partner in a timely fashion. Please give your partner at least a week to do the checks. If possible, send your soldiers by the middle of the month. Please take holidays into account when you do your planning.
• You and your partner must coordinate with each other. If you are going on vacation at the end of the month, let your partner know, so she/he can get you soldiers to check. You need to check those soldiers before you leave on vacation.
• If your partner does not send you soldiers to check, it is your responsibility to contact her/him, so you can get them. I have no way of knowing if your partner has not sent you soldiers.
• You must check your partner's soldiers in a timely fashion. Do not procrastinate and leave this task until the end of the month.
• Make sure you copy Sandy and me when you send the checking documents to your partner. We evaluate and use this information.
• If you have some really good reason why you don't think you can do checks during a given month (such as surgery, 3-week-long vacations), you need to contact me and make your case for why you should be exempt from checks that month. I will let you know if I approve the request. Please do this BEFORE I send out the month's checking partners, so I don't have make adjustments to the schedule. If you do not do this, I will assume that you've planned your work so that you'll be able to get your checking done.
• You must complete your checks before the end of the month and send them to your partner, Sandy, and me.

I recognize that checking is not any fun, but it is an important part of the work we do. We have found it useful for ongoing training. That is, we look at the checks and find things that inputters are forgetting or are confused about. This enables us to correct these problems and make sure everyone understands the rules. It also helps us ensure that the quality of our data is good. We would like to provide our users with the highest quality data possible.

Siblings living in the same household

Here is a post from Sandy in answer to a recent question:

After some recent checks, I would like to clarify the following guideline found in the manual:

When multiple siblings are living together...generally you'll probably want to select the lowest household number among the siblings living in the household.

As an example of this, if child Bob is household # 3 and his sister Judy is household # 7 and they end up living together one decade, then you should assign the family to household # 3.  This likely happens most of the time anyway if you input the households in order, which you should be doing.  (By nature of the fact that Bob is household #3, you would already have input Bob in all his decades including the one where he is living with Judy before you even start inputting Judy's household.) On occasion some NF's pop up later on when you do have a household for each sibling and you are forced to choose.  Since it's in the manual and we want to be consistent, we will  stick with the guideline of going with the lower household number rather than with the sibling who is the head of household as many of us have been doing.  Since it's really more of a guideline than a rule (it does say "probably" after all), and because it doesn't really affect the way the data user uses the data, it is not necessary to always go with the lower household number.  So don't go back to change it or tell inputters in their checks to go back and make changes.  But do try to go with the lower household number when possible.