31 March 2014

Correct vs Incorrect

I often pull topics for my blog from genealogy pages and groups. Whatever question is most often found for the week is the question I will try to answer. About a month ago, the question became "How do I do this right?" People were asking how to rate a source to determine it's accuracy, how to cite their sources, how to create research notes, how to make family charts.......... but it was all about doing it "right". Often, I answered the query as honestly as I could based on genealogical standards and personal preferences. Then someone else would come along and accuse me of being the Genealogy Police and saying I had no right to tell people how to handle their family history. Every. Single. Time.

Did they not notice the original question was asking how to do this right?

In the end, I came to the conclusion that I would devote more time to breaking down records and showing people what facts were available and how to source those records for themselves and others. Between the online conversations and a few tutoring opportunities (including teaching an aunt how to research for adoptees), I realised the new direction I wanted to take with my blog was going to be important for many who were just starting out or wanting to take genealogy more seriously. And that others would think that I was talking down to or otherwise insulting people "doing their own thing." So before I start my new format, I want to take a post to explain why there is a correct way of doing things and an incorrect way.

Now, in reality, there is only one truly incorrect way in genealogy: Clickophilia. If you start building your tree by turning off your brain, you're doing it wrong. Genealogy is "the study of generations" and does require your mental engagement in order to be successful. Whether that means you're going to draw your tree on the wall with crayons or write it out in a ten volume book set including glossy photos and detailed bibliography doesn't matter as long as you are willing to put on your thinking helmet. If your entire tree is made out of blindly clicking on leaf hints on Ancestry.com, uploading and merging other family member's Gedcom files, and entering in compiled genealogy books you find in the library (or more likely, Google Books), you are doing this wrong. So very wrong. But what is the "right" way? There is no *one* right way. There are no genealogy laws, so there are no Genealogy Police............. but there are genealogy software programs.

Now, I've said before, if you're going to have a paper research format, or keep your online work private, go ahead and add undocumented lines or links to Adam and Eve (or Odin or Hercules). You don't need to take research notes or write down your sources if you don't want to. I mean, you'll end up with loads of mistakes and no way to know if you looked at this record or that one, but that's your deal. When you collaborate with others, however, there is the unspoken agreement that you are both trying to be as accurate as possible. The other person won't grill you on your sources, but it's good to be prepared to give them as proof of your conclusions. Despite the connotations some would give it, a genealogist gets used to crafting a "Proof Argument". Basically, it's "I've found these records that prove my theory that X person lived in Y place and married Z person with these kids. I've also considered these records, but they can't be X person due to such and such reasonable doubt." The sources don't have to presented in MLA format and the argument doesn't have to be written out like a thesis. But when I talk about my grandfather being born in Morganfield Kentucky, I can "prove" that with his birth record, 1930 and 1940 U.S. censuses, World War II enlistment form, and SS5. I can argue that a similarly named fellow living at the same time in Arkansas is *not* my grandfather, because that man's death record doesn't have the same parents listed as the records I've found for my grandfather. That man's census records don't intersect with the rest of grandpa's family, and is miles away from any place he ever lived. But if I didn't have those sources, if I didn't craft a compelling argument, another cousin who only knew grandpa lived in Arkansas when he died may mistakenly believe this other man is our grandfather and start clicking away. They'd have to seriously shut off their brain, but it happens.

And that's the point: people have to shut off their brains to make a mistake and perpetuate it to others. Computers don't have brains. Remember the old adage "Garbage in, garbage out". If you're getting wild hints for an ancestor, it could be the algorithm pulling up records based on part but not all of your information in an attempt to find you records............. but it's more likely that a relative has connected that record to your ancestor in *their* tree and the computer system is giving weight to the record based on their attachment. Computer systems are built on the information that humans give them. Which is important to remember when crafting a search, too. How someone decided to index a record can affect how it shows in a search. If women are indexed with their maiden name, and you don't know it, having listed them in your tree with their married name won't help you. If an ancestor was born in, say, Orange County NC, but the part of the county he lived in became Caswell County by the time he died, you may be able to search for his death in Caswell and find it easily, but there won't be a birth record for him there! On the other hand, if you're using the system's mapping tools, those are based on today's landscape........ so listing his birthplace as Orange County will make it look like he moved when he didn't. Or if his town/county no longer exists, the map won't know where to pin him down. If you don't know a first name or a surname and you add characters like ? or * to indicate that, you could be using the search tool's wildcard characters by mistake and now come up with completely different results versus leaving the field blank. And while I'd like to agree with people who say that a couple that never married should never be listed as each other's "spouse", I've had unmarried couples identify each other as "spouse" in a record (for whatever purpose) and be indexed as such. Not having their profiles attached simply because I don't want to identify them as "spouse" means their names won't be included in a general search for each other and I may miss the very record I'm looking for. The computer system may ignore the record and not return it at all, or it may have it 10 pages in because it isn't weighted higher in the algorithm.

Search algorithms........ a computer programmer sat down and decided what fields would be included, and how close to the data input a record needs to be in order to be included. Without getting too technical, each field gets a score range and results are "weighted" based on score ranges. If I look for John Smith, I could get returns for John Cooper and Rob Smith just as much as I get John Smith. If I look for John Smith in Virginia, my first results should/could be for John Smith in Virginia, but could also be John Cooper in Virginia and John Smith in Maryland. If I look for John Smith born in 1940 Virginia, I get that and John Cooper born 1940 in Virginia, John Smith born 1900 in Virginia, John Smith born 1940 in Carolina............... with the most matched fields coming up to the top. And what I attach to my John Smith affects other John Smith searches (think of how Google will bring popular websites to the first page of search results). If a person is attached to me via Ancestry's Member Connect, that is taken into account by the system and the computer shows first what I've attached to our common ancestor. If 100 people attach the same photo to the same/similar John Smith, then that is ranked higher in the algorithm than the 1 person who attached the right photo to *my* John Smith (which I find on page 3 if I am willing to click through).

To better illustrate this: John Kemper of Fauquier County Virginia, born 1692 in Germany, died abt. 1758 in Virginia, married abt. 1716 to Ailsey Utterbach. Several hundred trees on Ancestry can be found. Hundreds of descendants connect to him. He's my favorite. His wife is alternatively listed as Alice or Alce, Otterback or Otterbach or Utterback. We all agree this is the same person and we attach the same familiar 9 records. One day...... or rather, one week, I decided to do a full search of Ancestry. The results were 70,000+ even with all the information I have available on him. Were there really only 9 real hits for this man in all of those 70,000? I made it about halfway through the results, I admit it. But I found 70 more entries for John Kemper than anyone else had attached to their tree. Many of them were genealogies of other surnames that connected to him via a child, but still. So now, when my aunt searches John Kemper with the same information, instead of the hundreds of descendants who couldn't think of a better name coming up first, she gets more of these lesser known hits. And the next cousin sees them higher up on the list because my aunt and I have attached them. And the cousin who's connected via Member Connect doesn't have to do a search because the system sends them an alert that we've attached the record first.............. and so on. But of those hundreds of trees, at least half of them have added census and land records for John Kemper II to his father John (I told you they had no imagination). So while the correct records are out there, census records are weighted very high and many will blindly connect to them. Now, those of you who have your thinking caps on are already aware that John Kemper d. 1758 couldn't possibly be listed in the first U.S. census of 1790. And while it may make sense to believe John had land, digging into the records would indicate he wasn't legally eligible to own land and records stating he lived on land owned by his son back this up. But the land records are only indexed online, so people see the name and the location and click click click............


So for the sake of the computer systems that know no better, there are some things we must consider a "rule":
- Date format: whether you put the month first or the day first, the month must be SPELLED OUT to avoid confusion. I've gotten so used to putting day first that in my everyday American life, I often confuse the dates my friends and coworkers write down (until the 13th day of a month, I lose my mind). If you don't know the exact date (or can't find documentation of a date for a fact) add Abt. for "about or around", Bef. for "before", and Aft. for "after". A baptismal record isn't proof of birth, so a birth year would be "qualified" by adding Abt. to the year or Bef. added to the date of the baptism. Same philosophy for burial records and censuses. If it doesn't directly describe an event, it isn't concrete proof of a date and qualification must be made.
- Locations: the old genealogy standard is to use the location name as it was at the time the record was made. Go ahead and keep that standard if you wish, but any location service provided by a computer will not put the results where they actually need to go. I use the current location in the field, and the original location in the description. I also put in my research notes a timeline of when the change happened. If you have FTM, you can add events to a general timeline that you can view for your person. It already has world events and will add family events (like the birth of children), so adding one for the change of a county name could help remind you.
- Maiden names: A woman should always be listed in a tree by her maiden name. ALWAYS. If you don't know a maiden name, I recommend leaving it blank. Some say dashes or underscores work for them ( - or _ ) without ruining the search algorithm and keeps them in some kind of order in their name index, and that's fine. But don't use any character that is a wildcard in the search tool! And LNU and MNU shouldn't be used as there are actual surnames Lnu and Mnu. Unknown shouldn't be added as the search tool will try to find that as a surname and when it can't, it will give extra weight to names that start with U or possibly sound like Unknown.
- Unknown first or last name: Like I said, no wildcard characters. If you don't know parents, but want to add siblings you do know, add a father profile with the surname but no first name. That's enough. You don't need to add a mother too, so there's no reason I should be getting tree hints for John being married to Unknown Unknown or my personal favorite: Mother MNU.
- Cite your sources!: When you attach a record from the search tool to your tree, Ancestry does the citing for you. Same with any tree service that also provides a search tool: you find the record and attach it to the profile, the company makes the link for you. But if you find records on one site and transfer the information to your tree on another site (or on paper/computer program if you aren't an online tree person), then you have to make some kind of citation for that source. WHY? First for yourself: if you know where you've looked, you won't look there again. And for others: you can save someone else a step. You don't have to put up images of a document you paid money for, or just hand over your research to complete strangers, but giving at least a citation of "found this death record on Missouri's Digital Heritage collection" could be enough to keep someone going. How detailed you make that citation is up to you (and I'll give you a couple of alternative citations in my upcoming records series).

There are more ways to create and maintain a family tree than there are flavors of ice cream at Baskin Robbins. But no one walks around with a lump of ice cream in their naked hand.
-Ana

No comments:

Post a Comment