work hours

I've noticed that some of you are consistently turning in fewer hours than you've committed to working. I understand that sometimes there is a holiday, so you miss a day or two. I understand that sometimes you are ill, so you miss a day or two. I understand that sometimes you take a vacation. That is fine. What I'm talking about is a pattern of working too few hours over multiple pay periods. Please evaluate your work effort and determine if you are working the number of hours your agreed to work. If you are not sure, please come talk to me. I track the number of hours you submit each pay period.

There are a few of you who have come to me and said that for certain reasons you need to reduce your hours. In general, I have approved those requests. If I've told you that you may work 15 hours per week instead of 20, then I'm not worried that you're not working 20 hours per week.

If something in your life changes, and you need to reduce your hours or take some time off, then please come discuss it with me. This is a job, and you are all adults. You have committed to doing good work for a certain number of hours per week. You are responsible to fulfill this commitment. If you are unable to do so, because of school or personal reasons, please talk to me, and we will work something out. If you just decide that working for us does not matter enough for you to fulfill your commitment, then we can discuss ending the relationship. If you do not communicate with me, then I am left to assume that you don't want to work for the project anymore. That is fine.

For those of you who receive health insurance through the Bureau, you must maintain an average of over 24 work hours per week. This average includes vacations and sick days (which will reduce your average). Our grant administrator calculated that if you work 26 hours per week, you should be able to miss about two weeks of work per year and still maintain your eligibility for health insurance. If you drop below the 24+ hours per week, you will not immediately lose your health insurance. It will, however, jeopardize your eligibility for health insurance next year. If you have questions about your eligibility for health insurance, please let me know. I'm tracking whether you're maintaining the hours for eligibility, but I do not know if you receive health insurance from the Bureau. If you are not currently receiving health insurance, but would like to make sure you're eligible next year, please come talk to me, and we can figure out if you're working enough hours.

Please let me know if you have questions or concerns.

Foster children and wards

There has been some confusion about the relationship code for foster children.

If the manuscript says the relationship is foster child, please use relationship code 5 for non-relatives. Do not confuse these children with those who are adopted. An adopted child should be relationship code 2.

Foster children are not the legal children of the parents. The situation is the same for wards. If the manuscript says the relationship is ward, please use relationship code 5.

Any questions?

Weekly update

Here are the stats showing how much work we did during the last week (through yesterday). You all did a really good job.

	Project 1	Project 2 - USCT
	n = 8,500	n = 4,500
Total soldiers complete	2503	369
Soldiers completed during last week	47	46
Soldiers w/o children (to be removed from sample)	276	87
Soldiers with children complete	2227	282
% with children complete	26.2	6.3

Reminder and question

Reminder

Just a reminder that speed counts (unless you are training). If we are too slow, we won't finish the work. There are a few people who are considerably slower than other people even when taking into account that people work a different number of hours and some families are harder than others.

Since you keep track of your times, please calculate your average time per soldier and make sure you are going fast enough. If you are not, please concentrate on speeding up. Those of you who are quick, feel free to share with everyone any tips you have for being both quick and accurate.

Question

Do you have any questions you'd like answered or topics you'd like discussed on this blog? Please let me know.

Reminder - weekly updates

Please remember, that if you don't work in the office I need a weekly progress report from you. I need it by Wednesday at noon eastern time. Every Wednesday.

You may either email me your update or you may share your input log with me. Let me know how many soldiers you've completed and how many did not have children.

Some reminders about Project 2

There has been a little bit of confusion about what you're supposed to do with HH0 for your Project 2_USCT soldiers, so here is a short review.

• You may spot check the original work to help familiarize yourself with the soldier and his family.
• Move the URLs from the remarks to he correct field.
• Correct M1/M2.
• Correct any obvious mistakes.
• Do not spend time searching for the soldier's NFs. We already had an inputter spend a lot of time on this.
• If correct decades come up as hints in your tree, please input them.
• If the original work is wrong, you may delete it (or ask me to delete it), but there is still no need to go searching for him.

When checking Project 2 soldiers:

• Do not do anything other than spot check HH0, and only do this in the most cursory manner. Your time is better spent on the children.
• Do not search for the NFs for HH0.

Any questions?

Notes from the Census Inputters meeting February 10, 2016

Here are the notes from our recent meeting. Please read them and let us know if you have questions or concerns. Sorry, tables and images did not copy correctly. Please refer to the emailed notes for the missing images.

ANCESTRY ISSUES:

Family Tree:

·         You do not have to clear all your hints nor do you have to attach all the confirmed hints to your person.  Often numerous military records pop up in hints that we already have collected through the pension and military service collections.  You do not need to take the time to attach all these hints. ALWAYS attach confirmed WWI and WWII draft cards.

·         You do not have to click “Ignore” for all the bad hints either if you don’t want to.  You can leave them hanging.

Searching Notes:

·         Searching by Race: If you want to narrow your search to only black people and exclude whites, you will have to do separate searches for Negro, Black, Colored and Mulatto.  See the chart below for the predominant term used in each decade.  Typing in one of those will not find the others.

I have found if black is somehow added to the index in brackets (not parenthesis), then a person labeled originally as Negro will appear with a search of black/exact. However, if a Negro has black on the preview page in parenthesis (not brackets), that person will not be found with a search of black/exact.

If you put in black but don’t mark exact, you’ll get results that include whites, but all the black people will be listed first.

  

·         Searching with Family Members: Ancestry has changed the nature of searching with family relationships.  Previously, you could add children’s names without it being an exact search.  Be aware that now it is an exact search for whichever children’s names are listed despite the fact that there is no check box marked “Exact.” Hopefully Ancestry will fix this.

·         Searching by Street: You can search with street name on 1910, 1920 and 1940.  If you get some good finds using street names, send Sandy an email with the recid, the decade and the person found. 

Source Citation Changes

·         Page numbers aren’t showing up in the source citations anymore. We don’t know if this is temporary or not.  See the info under Page Numbers below to determine what to do for these cases.

INPUTTING ISSUES

Page Numbers

·         Sometimes there is more than one page number listed on pages in the 1850-1880 census manuscripts.  Sometimes you have to go the previous page.  Ideally, the page number on one of those two pages will match source citation.  However, occasionally there is a discrepancy.  And currently the source citation is not showing the page number.  Here are the rules for taking page numbers in 1850 – 1880: 

o   Use the page number from the source citation.  If there isn’t one, then…

o   Take the number that is stamped on the page. 

o   If the page is blank, take the number that is stamped on the previous page.

o   If there are two handwritten numbers, take the one that is NOT next to the words “Page Number___.”   You might have to turn to the previous page to find it.

o   If the only page number you can see on your manuscript page and the previous page is the one written in next to the words, “Page Number__,” then take that number.

Children on grid

·      

            Be sure to put dead children on your grid.  For unnamed dead infants in mil info, input them as:

o   Last name, (dead infant)

o   Mark not searched across the row

·         If it’s just a child with no birth or death info, you will have to input as Last Name, (Child), and mark him not searched in any decades prior to the vet’s marriage date and yellow for all other decades. 

·         Remember for Project 2, you can mark all children in 1850 and 1860 as not searched. 

·         Do not use the 1900 and 1910 census information for living and dead children as the sole source to add unknown children to the grid.

·         Use your best judgment and add children to grid if there is evidence of additional children even if you don’t have names or birth/death dates. 

·         FYI:  The 1898 family circular did not ask for dead children, but the 1915 circular did specifically ask for dead children. 

Adopted vs. foster or wards

·         Adopted children are 2 on manuscript with note in remarks that it is adopted child [Relation: adopted child].  Search for adopted children.

·         Foster and ward should be marked 5 (non-relative) in the screens and non-relative in update relations.  Do not search for foster children and wards (unless their status changes to adopted child or child)

Step Children

·         Search for step children if you determine that they lived with the veteran for a number of years during childhood, defined as years prior to age 15.  Be sure to calculate the age when they moved in with the veteran, not just the age they are when you find them in the census with the veteran.  (If you find a 17 year old step child with the vet in 1900, but you know he married his spouse in 1890, you can assume the step child has lived with him for 10 years since age 7 and that you therefore need to search for him in other decades)

·         If several step children live with the vet, still go by the age rule above.  For example if three step children move in with the vet at ages 17, 13 and 10, you would only have to look for the 13 and 10 year olds in subsequent decades and not the 17 year old.

·         Step children, if labeled step on the manuscript, are entered in screens as a “0.” However, be sure to input them on the update relations page as children of their birth parent, not as step children.

·         Remember you do not have to search for the mother of the vet’s step children if she doesn’t give birth to any children with the vet.

MilInf? Column

·         Mark this column for anyone you find in the census that is in mil info.  This includes parents, siblings and people listed in the “Living With” tab.

Death Ranges

·         Even if the dead by date is on the mil info, the death range in the obit still has to be a QC 3 (not a QC 1 for mil info death information) because all death ranges are QC3.

·         You do not need to make a note of the month when doing a death range from the dead by date in mil info.  (This date was the date of the family circular document.)

Out of Country

·         If there is evidence from the mil info that your vet has moved out of the country and does not return, you do not have to continue searching for him.  If he moves with his family out of the country, you do not have to look for his children, but do search for their death dates.  Mark them not searched for decades when they are living outside the US.

PROJECT 2 ISSUES

Here a few reasons that Project 2 is a more challenging sample than Project 1.

·         Veterans have lower quality codes to begin with, some 3’s and 4’s even.

·         Children aren’t always listed in the mil info

·         There are not as many exact birth dates in mil info and often no birth dates at all

·         When birth dates are in the mil info, they don't always seem very accurate.  Also, the ages given in the census aren't very accurate either.   

·         There are more children in mil info listed just as child with no names than there seemed to be in the white sample. 

·         Veterans may have more than one name, sometimes a slave name then a new name or another alias they go by.

·         Person searched for is often not  head of household but might be living with his spouse as servants in someone else's household which makes it harder to search using family relationships

·         Sometimes the married couple is not living together (living in separate households as servants or just living separately)

·         There are often blurred lines on marriage, living together, who’s the mother, over lapping spouse dates, multiple spouses. [Example: Married in 1886 but a couple of children born before that. (Fausky 9000107141); mil info says married in 1899 but found as single boarder in 1900 with the eventual spouse(Marshall, 9000107176)]

·         Very few hints are showing up because there aren’t many family trees for these families on Ancestry

  

·         Very few Find A Grave records

·         It’s harder to use 1900 and 1910 census info (from number of living children/number born) to determine estimated death ranges because it’s harder to trace the living children.    

Tips to keep in mind when doing Project 2:

·         Don’t forget to put in M1 or M2 in 1910 for the vet.

·          Even though we are not checking the previous work carefully, it’s OK to build your tree with the veteran’s censuses.

·         Understand that you will have a lot more yellow on your grid than you are used to.  That’s OK. 

·         You’re just not going to find a lot of them, so give up sooner.  If you don’t have enough to go on and you can’t find any records, it’s ok to give up.

·         Be cautious of taking a guy in 1870 because he’s the only guy that fits when you have 12 guys the next decade that fit.  Where were the other 11 in 1870?  They may just not have been enumerated.

·         Because birth dates tend to be less accurate, use at least a 5 year window  or even 10 for the birth age (or consider searching without age sometimes)

·         There is a wide variation in spelling and names, so be creative in your searches. 

·         Don’t get suckered in to thinking you’ve found marriage records for your families prior to 1860 in the south.  The names may match, but because race isn’t listed, you may have found others with similar names, likely white. There weren’t really any decent records in the south for blacks prior to 1865.

·         Quality codes for the vet are based on what is in the mil info.  You may find your vet with children not listed in the mil info and he’s only a QC3.  When you find those kids in their own households later on, then they can only be QC3.  We are using the QC to say how certain we are that these are the children of the actual soldier in the mil info, NOT how sure we are that they are children of the person we found in the census data which may or may not be the veteran. 

However, if you find a good death source that confirms your QC 3 guys is actually your guy…..(for example, there is a name of child not found on mil info but on a death cert that is living with your vet in the QC3 decade, and the death cert matches the death date on the mil info), then you can make the obit a QC1.  The decades still have to have QC’s based on mil info only, so it would remain a QC3.  Now that the obit positively confirms that the soldier from the QC3 decade is really our guy, you can use it (the obit info) to increase the quality codes for the children when you find them in their own households.   In other words, you can use other confirmed sources to bump up your QC for the children, but the vet’s QC has to go by mil info.

·         Don’t get discouraged!  If you need a break from Project 2 guys for a while, intersperse your Project 2 guys with some from Project 1.

Comments that came out of meeting:

·         Input what you see in the manuscript even if it’s wrong.  Make sure the Update Relations page reflects the accurate relationship.  (So if it says head, wife, son, wife, input what you see even though you know the 2^nd wife is really the wife of the son and not the wife of the head of household) 

·         However, keep the egregious error rule for sex and age.  (Example 1: If Maria Susana says male or a daughter who is male, then leave the sex blank and note in remarks that it says M but should be F. Example 2: If the census says the head of household, who is a farmer, is 4 years old, leave the age blank and make a note in remarks that the age listed was 4 and thus inaccurate.

·         If the supervisor’s district (or any other information) is crossed off, don’t input it.  It’s ok to be blank if there’s no information to replace what has been crossed out. 

·         For families split on two different pages, take the page number that links to the URL and the person you searched for. 

·         Sometimes two dates are listed on the page under date of enumeration, and you can see written in the margin where they started on the 2^nd day.  Traditionally we have taken the first day.   However, it is not wrong to take the second date.  You do not need to remark on it in you checks.

·         Make sure Heather gets all your completed trees and that they’re not getting lost.  The trees are important for what Chris is working on.  He is trying to adjust the screens and create a GEDCOM uploader that takes data from the trees to automatically fill the screens and save us time in inputting.

·         Shared trees will disappear from your shared list if the owner of the tree deletes them from Ancestry. You can delete trees yourself by going to Manage Trees, then Trees Shared With Me and clicking Remove from list

  

·         Don’t rule out white people when searching for Project 2 people.  Open the manuscript to look at it.  Occasionally we have found that Mulattos are indexed as W even though it clearly says M. Some WWI and WWII draft cards show blacks listed as white. 

·         Don’t spend a lot time doublechecking vet’s work in Project 2.  Make sure the right decades are chosen. No finds have been checked and re checked.   If you find someone better, by all means replace the old data. 

·         For obits if only the county is listed, and you’re positive what the state is, you can include the state. 

 

·         To get the source citation page number, click on “view record” to get to the preview page, not the manuscript image, and look for the source citation information in the fine print at the bottom.

Social Security Indexes

If you come across a Social Security number for the soldier, his spouse, or his children ALWAYS input it.

If your death source is the Social Security Death Index, select that as the source and input the number in the box that appears.

If you find the U.S. Social Security Applications and Claims Index, 1936-2007, enter the Social Security number in the Remarks field of the appropriate individual's Death screen. Type SSN and the number.

We want to make sure we ALWAYS collect the Social Security number.

January 2016 Checking Stats

In January, we checked 42 soldiers under our new 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? - 5
• Missing HH member - 11
• Duplicate people - 1
• Wrong person - 1

Inferred Relationships

• Incorrect relationships - 16

Census Errors

• Name - 44
• Typo/Reading/Wrong - 122
• State Code - 0
• Missing/Wrong URL - 5
• Missing data - 60
• Additional finds - 13
• Quality Code - 19

Death Errors

• Typo/Reading/Wrong - 9
• Missing data - 15
• Missing/Wrong URL/Source - 14
• Quality Code - 7
• Additional finds - 13

Tree Errors

• Missing/Incorrect information/relationships - 14

The total number of errors for all 42 soldiers is 369. Some differences/errors affect the data more than others. If we checked other pensions, we'd probably find similar differences. The inputter whose checks were the best, with a total of about 3 differences, was Irene.

More on soldier's dead children

If the Mil Info says the soldier has a dead infant, we want that child added to the Grid. If you think this is the same as another child you've already added to the Grid, then there is no need to add the infant a second time. Sometimes you won't have any information on this child - no name, no birth date, no death date, or living/dead date. In this case, please do the following.

Add the child to the Grid in the following manner.

Jones, (DEAD INFANT)

Mark all the Xs black to indicate not searched.

We can discuss this further at our upcoming meeting.

Clarification on soldier's dead children

If you have evidence that the soldier had a child, make sure that child is entered on the Grid. If the 1910 census says the mom had six kids and four are alive, you don't automatically have to assume that the soldier had two more children that you didn't know about. But if your searching turns up evidence of children, even though you never find them in the census, please add them to the Grid. If the Mil Info says the soldier has a dead child, and you never find any children in the census, you still need to add that child to the Grid. This is the same as our rule for stillborn children.

Our goal is to account for all the soldier's children.

Dead children of the soldier

Sometimes the Mil Info or another good source says the soldier had one or more dead children, but doesn't give any names. A child is a child whether alive and named or dead and unnamed. All these children must be added to the Grid. Add them in the following manner.

Last name, (Child)
Smith, (Child)

If you don't have any birth or death dates, use what information you know to infer. You might have a dead by date from the Family Circular. Or the child is listed as dead on a census. Fill in black and yellow Xs as appropriate.

We just want to make sure that ALL of the soldier's children are added to the Grid. If you think this dead child refers to another child you've already entered, then there is no need to enter the child again.

On checking

Here are some notes from Sandy about reviewing checks that you receive.

It is important to review checks carefully, not just look at things you need to change.  Really see the mistakes and understand what the problem was.

If there are any matches you didn't find, go back and look at it and repeat the search the checker did. See if you get the same results and try to understand why the checker's search worked.

Open the manuscript occasionally and look at handwriting that you couldn't read the first time that the checker was able to figure out.  This helps a lot in getting used to difficult writing.

If there is anything you don't agree with, then ask questions. This is an excellent way to make sure everyone is working with the same rules.

We'd like you all to benefit from the checks. If you do anything when reviewing your checks that helps you be a better inputter, please let us know in the comments.

Project 2 - USCT

If you've been working on a Project 2 - USCT assignment, I would like feedback from you.

Here are some things for you to consider:

• How does it compare to Project 1?
• Do your soldiers have a lot of kids or very few?
• What are your overall impressions of the work?
• Anything you think I might find useful in evaluating the project

I've noticed that some of you are burning through USCT assignments really fast. Others of you are moving very slowly. Why is this? Are there some assignments for which the soldiers have lots of kids? How are you at giving up? As experienced inputters, you know when you have enough information to make a match.

We've noticed that sometimes the Mil Info is not uploading correctly to the Project 2 database. If you notice that the extant data has QC 1s and 2s that don't make sense based on what you see in the Mil Info, please stop work on the soldier and email the name and RecId. I'll be able to figure out what you should do with that data.

Any feedback you can provide on Project 2 will help us all be better at collecting this sample.