Friday, July 31, 2015

Lesson Learned Again: Don't Trust Online Family Trees

Well, at least for colonial ancestors who migrated from England to New England during the 1620-1650 time frame.

I wrote Lesson Learned: Don't Trust Online Family Trees on 24 April 2012, saying:

I received an email this week about the ancestry of Susanna Page (ca 1611-1691) who married Thomas Gleason (ca 1610-1686), asking if I knew who her parents were, since there were several sets of parents given in the Ancestry Member Trees. 

The background on this can be found in:

*  Who were parents of Susanna Page (ca1611-1691)? (posted 19 April 2010)

*  English Parish Record data and Forrest Gump

*  Images of the Will of Thomas Page (died Hawstead, Suffolk, 1637)


To summarize - I found that Thomas Page of Hawstead, Suffolk was the father of Susanna (Page) Gleason, based on a message board post that referred to a 1637 will in the Suffolk Probate records.  While in Salt Lake City in 2010, I found the 1637 will of Thomas Page of Hawstead with the help of the FHL staff.  

Today, the thought occurred to me - has the percentage of persons with the correct parents of Susanna (Page) Gleason (ca 1611-1691) improved from the 0% three years ago?

For the study, I used Ancestry Member Trees, and these search criteria:

*  First Name = susan* (exact)
*  Last Name = Page (exact)
*  Birth year:  1614 (plus/minus 5 years)
*  Spouse first name:  thomas (exact)
*  spouse last name = gle*son (exact)

There were 979 matches for these search criteria:


I then added more search terms for different names of fathers and mothers I saw in the list of matches:

For instance:

*  Father's First name = john [exact] = 320 matches
*  Father's First name = john [exact] and Mother's first name = ph*be = 315 matches

*  Father's First name = robert [exact] = 56 matches
*  Father's First name = robert [exact] and Mother's first name = lucy = 15 matches

*  Father's First name = william [exact] = 14 matches

*  Father's First name = thomas [exact] = 61 matches
*  Father's First name = thomas [exact] and Mother's first name = susan* = 10 matches

Here is screen shot of the last, and correct, search results:



So 10 out of the 979 matches for my Susanna Page provide the correct names for the parents - that's 1%!!  That's up from 0% three years ago.

The 61 matches for the father being Thomas means that over 6% have the right father, which is up from about 1% in 2012.  

Kudos to those 1% that have added the correct names as Susanna Page's parents - it's taken 8 years to get to 1%!  The first online record of the correct parents names were published in  GenForum message written by Dan Page on the GLEASON Message Board on 19 August 2007.

I didn't look at other family tree resources, but my guess is that the results would be similar.  The exceptions might be the collaborative trees - Geni.com (NO), WikiTree (YES!) and FamilySearch Family Tree (YES, but several duplicates, including many wrong parents).

My comments from three years ago still apply, I think:

What does this mean?  As we all realize eventually, many of the entries in online family trees (including Ancestry Member Trees) are copied from the trees of other researchers without any check of the possible records involved.  I admit that I did this over 15 years...and now I'm in the process of trying to fix those problems by finding authoritative sources (published, images on microfilm, or images in record collections). That's a lot harder to do than copying information from an online tree, but it's the right thing to do.

The lessons learned (again) is:  Don't trust online family trees.  Use them for clues.  Verify the assertions, add source citations for assertions, etc. for every person in your tree.

This problem seems especially common with the immigrants to North America during the early colonial years.  There are thousands of persons whose ancestry is not known, or known erroneously, from this time period because of the lack of research in original records.  For early colonial New England ancestry, research in the Great Migration volumes (available on AmericanAncestors.org and Ancestry.com) may be helpful.  The peer-reviewed journals, like the New England Historic Genealogical Register, the National Genealogical Society Quarterly, The American Genealogist, and others frequently have articles about the ancestry of early colonials.  


Copyright (c) 2015, Randall J. Seaver

Please comment on this post on the website by clicking the URL above and then the "Comments" link at the bottom of each post.  Or contact me by email at randy.seaver@gmail.com.

10 comments:

Linda Stufflebean said...

Randy, I did an experiment with my husband's John Whitmer at the beginning of this year. Of 350+ online trees, only a handful had the correct parents. I emailed everyone with a polite message about research proving the parentage. I got replies from fewer than 30 and, of those, again only a handful updated their trees. It's not very encouraging with your findings, either.

Janice Harshbarger said...

I think a large part of the problem with the 99% is that they did their "research" once and then have never gone back to see what's been learned since then. I plead guilty to that charge. Honestly, we are never going to be able to go back to check every single person on a tree of tens of thousands, but I find that writing blog posts, and checking those folks out in the light of what I've learned since I started family history, helps a lot. I've corrected errors that way, and learned more about almost every ancestor I've written about. It's a good feeling to correct a mistake, almost as good as finding someone new!

joanne malene said...

Ancestry gave me a shaky leaf today on someone who could be my 6x second cousin - there were 12 ancestry trees with this cousin - EVERY ONE of the trees had the exact same father and mother for this girl. The problem was, the parents were already dead when she was born and had been dead for around 100 years. I emailed all 12 and just heard from one who told me to mind my own business... well, o.k. then!! :)

Suzette Frederick said...

Early in my genealogy I did the same thing, taking information from others trees. Now correcting my data and sources. Sure is easier if you do it right the first time. With several couples last night and a couple want to start their genealogy. I ask if they'd like to get together so they could learn from my many mistakes. I hate that we exponentially share bad data

David Smith said...

I think I've got the record

John " Rock" Smith (abt 1615-abt 1706)

Almost 5000 entries on Ancestry and all but about 10 of them have multiple errors.

Neither the names of his parents or wife are recorded anywhere yet almost every tree has names recorded based either on misinformation or wild speculation.

There are three contemporaneous documents that bear the names of his 6 children, but that doesn't stop people from adding anyone they want to connect.

Most of the entries literally have every possible error recorded in them.

Russ Worthington said...

Randy,

I'd like to do an experiment on this topic, or you can.

See what happens, if you post a small GEDCOM on an Ancestry Member Tree and, withOUT searching, what are the Online Tree Shaky Leaf Hints provide you.

I find that relying on Shaky Leaf hints, in many cases, return a better result list then if I manually enter a search as you did. I know that the Record Hints are very good, not perfect, but very good (on my main tree between 90 and 95%) good results.

Not questioning your results at all. I totally get that, which is why I don't spend a lot of time on online trees.

We we had a way to search on a specific USER's online tree, or have the ability to compare trees like we to with Ancestry DNA.

Thank you,

Russ

Kevin Sanderson said...

Hi Randy,

Good stuff! Thank you for posting Jonathan Sanderson's last will a while back!

I would also be wary of the so-called peer reviewed publications. It seems some submitted articles without verifying information with the original records. You will run into this regarding Edward Sanderson and William Sanderson taking the oath of fidelity. One article had the year wrong, and another had the wrong person. The correct answer based on the original and the extracted records of the Middlesex County Court records has the oath being taken by Edw. Sanderson in 1652 (FHL film 892250). People have had both taking the oath or just one or the other with the wrong year for over 100 years since the articles were first published. William Sanderson if he was born in 1641 based on one later deposition, would have been too young, so it has always been a source of some debate. But now we know it was Edward in that list of names. (We also now know William was Robert's nephew from a later item in Robert's will, and yDNA testing (FamilyTree DNA) so far proves William is related, but not Edward's son, so the son of another brother of Robert.)

The same caution goes for the work of Dr. Bond. He did a good job gathering together much information about the families in Watertown, but he makes some conclusions and assumptions that are wrong. I think part of that was from the information he was paying people to collect for him, and even he has a disclaimer saying some may not be correct, something amateurs always miss it seems. Many quote him saying there's no affinity found between Edward Sanderson and Robert Sanderson. Well, Edward is mentioned twice in Robert's will as his brother. Edward also sold land east of Robert's land to William Shattuck, Sr. in 1664. We know about that land deal from Shattuck's will and the Selectmen weighing in on it though the deed is missing (can't be located in the records so far, but many allegedly never were recorded), but a later sale of part of the the land by William Shattuck, Jr. to two Livermore women places it where John Livermore had earlier owned swamp land and it's placed east of the heirs of Robert Sanderson's properties in the deed and south of William Shattuck's other properties near Fresh Pond. The younger Livermore family members had been buying much of the old holdings, plus some, in the area. I speculate that John Livermore sold part of his swamp land to Edward Sanderson when Livermore moved back to Watertown around 1653, but that deed can't be found, though we know the deal happened with William Shattuck and the family history of Shattuck's nice view of Fresh Pond and Cambridge and Boston in the distance (a view not provided from the Clough property where he lived for several years - he lived on Edward's old property for 8 years until he died).

Page Sanderson copied many of Bond's mistakes along with others in an article she did for the New England Historical and Genealogical Register in 1973 (vol. 127).

Yep, tread carefully!

Kevin
http://www.ksanderson.com/edward_sanderson_of_watertown_ma.html

Kevin Sanderson said...

Here's more on Page's mistakes I couldn't fit in on the previous comment...

Here are some of the places Page Sanderson was mistaken. The first error is the birth date and death for Edward Sanderson... we are not completely sure yet that Edward's birth was 1611, in Higham Ferrers (Robert said he was from Higham when registering as an apprentice in London, and there are a few other connections that point to Higham Ferrers - not Norfolk, Norwich or the Higham near there, but it sure looks like Edward Saundersonne in the original parish register, not Edmond Saundersonne that Rushden Research has insisted for years) but it was often different in many trees and histories that use 1614 or 1615 like she does. The death date of 1680 can't be correct as he was mentioned in deeds and wills through 1696 - Robert's will was 1693. She mentions the often reported error that he came over on the Increase, but correctly mentions that he wasn't on the passenger list (Robert was not either, which you will often see). There are no records of Edward's and his wife Mary's deaths in Watertown that have been turned up yet.

Page has Edward moving to Cambridge in 1654 saying he died there in 1680 (and mentions another source saying 1674 as his death year), but our Edward sold his property to William Shattuck in 1664 and then moved to western Watertown according to genealogist/historian, and one of the founders of the Waltham Historical Society, Edmund L. Sanderson, who lived in one of the old Sanderson homes from the late 1800s on.

Another error is saying Edward and Robert are not proven as brothers, though she does mention he's in Robert's will. The will is proof! Even if there weren't the land connections I have found, the will is proof!

She has the wrong year for Edward taking the oath which means she also didn't look at the original records.

Page does not give a source for Hester's birth and many people copy it. Benjamin Worcester in "The Sanderson Homes at Piety Corner, Waltham" at least said there was some question about Hester being a child. I've looked for Hester in the Watertown Vital Records and her birth is not listed. Edmund L. Sanderson completely avoided mentioning Hester in his article about the Sanderson family for the society. Jonathan is the only child he mentions saying he was the only son to reach maturity and there were other children, but he doesn't state how he knows that, other than Jonathan being the only recorded child.

Page gets Abigail from Bond, but she lucks out there as Abigail Saunders married Shubael Child and the Child family had land south of Edward in western Watertown (John Ball sold woodland to Richard Child south of Edward in 1676). She appears likely to be a daughter from the location information. If you adjust her often mentioned birth year from 1653 to 1663, she could have been the older daughter who was apprenticed out by the Selectmen when Edward was having a rough time apparently health wise and financially (his financial problems go away in 1676 as the town no longer records any reimbursemnt to John Flagg for an apprenticed daughter of Edward - I speculate that Robert finally learned of his troubles then - and it might explain why Jonathan had a lot of money when he died and why Jonathan was not mentioned in Robert's will - professional genealogists say if someone wasn't in a will, they probably already got money or help years before).

She puts William's daughter Hannah in as Edward's child, using the wrong birth date... ignoring the Julian calendar, like Bond did. So she got that from Bond.

She puts Ann in but we know she was John Saunders' daughter from the First Church in Boston records. She got Ann from Bond, too.

On page 184 she does get it correct about the clerk errors!

The rest appears to match the existing records. It's the early stuff that causes problems and headaches.

Kevin

Kevin Sanderson said...

There's also the mixup of Edward Sanderson with a poor commoner named Edward Sanders/Saunders who lived on King's Common near the Hugh Parsons family (we know that from Ruth Parsons' General Court testimony and we know where the Parsons family lived, on King's Common on the Cambridge border). Edward Sanders often received town aid, usually from the deacons living near King's Common being reimbursed. Hester Sanders could have been one of his many children.
Edward Sanderson isn't recorded needing town help until his three old neighbors from eastern Watertown, John Coolidge, Thomas Hastings, and William Bond were reimbursed for giving him some corn in 1669.

Sue said...

I have always said to people who contacted me, use the information but please double check as I have may have gone wrong somewhere. I one case some did contact me to say I had gone wrong, but could not work out how and between us we found the wrong. But a second person wrote a very nasty email to me to say I had gone wrong and I should change it - I wasn't very nice back as the information was about my Grand parent's and my mother, who was in her 80's at the time was the one giving me the information. So sometimes the people who are asking you to change the mistakes are sometimes wrong too!! So in the end check your own information and do not rely on others. Because of this I have taken to: 1, not letting every one see my trees and 2, putting messages on parts of them such as 'this part has not been double checked' or 'date put in for search purposes'