Jump to content


Photo

Accept or Reject a Webhint from an Ancestry Other Member Tree


  • Please log in to reply
4 replies to this topic

#1 Roberto

Roberto

    Member

  • Members
  • PipPip
  • 20 posts

Posted 06 July 2019 - 06:24 AM

I acknowledge that this is more of a methodology question but it arises from RMs Webhints.

 

Webhints provides me with a number of Ancestry Other Member Trees that include my great grandfather and his family in the early-1800s to early-1900s.  When reviewing them, I find that most show details of him, his wife and their ten children.  Generally all information is correct, accordingly to my evidence, except the birth place of him and their first child. Great grandfather shows a location in England whereas he was born in Kingston, Jamaica and for first child, shows Kingston, Jamaica instead of a town in Australia.  (I have good evidence for both dates and places for both of them.)

 

My expectation is that I Reject the Webhint for my great grandfather and for his first child, even though the majority of the family's data in the Other Member Tree is correct.

 

Will my Rejection then limit their inclusion in further potentially useful Webhints?

 

Am I using Webhints as expected by Bruce and his design team?

 

Cheers

Robert



#2 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3955 posts

Posted 06 July 2019 - 06:55 AM

You ask a hard question. I run into the problem all the time of 90% to 95% of the data included in a particular Webhint is correct and the rest is incorrect. What is one to do?

 

First of all, your question is specific to ancestry's WebHints. If instead it were for FamilySearch's WebHints, I'm not sure that my answer would be exactly the same because the hints work very differently for the two systems. Because of the specificity of your question, I will answer only for ancestry.

 

The first thing to understand is that strictly speaking, there is no such thing as WebHints in RM for ancestry. What there is are shaky leaves in ancestry. You can manually type in a new tree into ancestry and as soon as you get far enough into the past with your tree to start matching with other people on the ancestry site, some of the people you typed in will start showing shaky leaves. You can approve or reject these shaky leaves. These shaky leaves are hints that can be 100% right, 0% right, or somewhere in between.

 

With RM's TreeShare, it creates your tree on ancestry for you based on what is already in RM. Or if you already have a tree on ancestry, it copies the tree for you to a new and empty RM database. Either way, RM shows you the shaky leaves on ancestry within the RM user interface and it calls them WebHints. RM doesn't have any WebHints of its own.

 

That being said, I tend to want to process and dismiss the hints when I see them, rather than leaving them dangling forever. The only two ways to do so are to approve the hint or reject the hint. This is just me and it may not be right for you, but I tend to approve such hints if they clearly are the correct person and even if the data is not is not 100% correct. I will often make attempts to contact the person with the incorrect data to notify them the situation.

 

But I never, never, never bring any data from WebHints into RM automatically.  I type everything into RM manually. So my approval of an ancestry WebHint does not have any adverse impact on my RM database. My rationale for this approach is that I can see an approved hint again if I want to do so, and the best I can tell I can never see a rejected hint again. There is no way that I know of to get it back. If you use TreeShare to bring data back into RM automatically, you surely would not want to being any incorrect data back into RM automatically, so my approach might not work for you. You might want to reject the partially incorrect WebHints.

 

This is just my approach. It is surely not the only reasonable approach and it surely won't meet the needs of everybody. But it does seem to work for me. I'll end by saying that in reality, the ancestry WebHints that are for other trees are much less reliable and much less important than the ancestry WebHints for records. Those records are what you really want. I sometimes wish TreeShare had an option to show only the hints for records and not to show the hints for other trees. But ancestry really wants you to hook up with the trees because (as their TV commercials like to say), that's the way you can complete your family history very easily in only a few minutes (even it's completely wrong!). I find the hints for trees to be a good way to hook up with other researchers who are working on the same families that I am, even if I don't blindly copy their data.

 

Jerry



#3 Rooty

Rooty

    Advanced Member

  • Members
  • PipPipPip
  • 267 posts

Posted 06 July 2019 - 03:52 PM

Jerry: that was a very clear summary of how to use web hints data and a caution on the auto merge process. I printed your comment out to think it over carefully.

 

FYI, I checked and my family tree maker has a preference to see ancestry and family search hints (if you are logged in) and another checkbox to NOT see ancestry tree hints. That reduces clutter greatly but I have found leads worth pursuing even in trees with bad data (most). You just have to check them out yourself. Large trees that reach back to the first neanderthal ancestor really make you wonder about people's judgement of data quality.



#4 Kamolga

Kamolga

    Advanced Member

  • Members
  • PipPipPip
  • 70 posts

Posted 07 July 2019 - 12:58 AM

 

But I never, never, never bring any data from WebHints into RM automatically.  I type everything into RM manually.

This is just me and it may not be right for you, but I tend to approve such hints if they clearly are the correct person and even if the data is not is not 100% correct. I will often make attempts to contact the person with the incorrect data to notify them the situation.

ancestry WebHints that are for other trees are much less reliable and much less important than the ancestry WebHints for records

100% identical for me and I frequently ask myself the same question.

 

First, manual input=consistency=database query with perfect output. Therefore automatic importation is a huge loss of time in my way to manage my database (checks, corrections and my database would make no more sense to me)....but it can work for other people, I just can not combine it with my input and still save me time. I tried to change my input but it did not work.

 

I basically have very little interest in shared trees (errors are copy-pasted from one to another) but for the edges of the branches to find insights...and for those I would rather look for the person on shared trees websites directly if I can not find better records. If you opened my RM file, you would see a single database but in my mind it is split into proven relationships and facts vs shared trees origin, which I recognise by putting an * in the suffix and about (abt) in the event dates -> I wanted a way to see it without opening anything like sources or notes and the computers to still be able to match the people with other databases. I even use a SQL query to automatically write or erase the legend of * and about at the end of the note sections (person note for *, fact note for abt) so that when I share my tree, other people know that those are not reliable.

 

I also found plenty of official papers with mistakes, spelling in names (many ancestors were Belgian illiterate farmers)...I even found some papers where I am Syrian (I was born Belgian and never got Syrian nationality but my father was) and my own father with 3 different birthdate (that can be explained by the fact that none of his brothers know his exact birthday, they roughly know the period of the year but he had chosen only one). In your case, I guess when you input Kingston in a software, the first that comes out is Kingston near Wimbledon in Surrey, South West London (I lived and studied there :) ) but I think it is common interest to recognise the person is the same person even if some mistakes, so the logarithms work better, with more matches and more users can access your findings. I wish RM had its own tree sharing system where sources would be more visible than with their partners.


Rootsmagic 7.5.9.0 with a lot of SQL queries (SQLiteSpy) and a bit of Family Historian 6.2 (tree view and map)


#5 Roberto

Roberto

    Member

  • Members
  • PipPip
  • 20 posts

Posted 08 July 2019 - 11:50 AM

Thanks everyone for the thoughts.

 

I also don't bring in any data from external sources but type it in myself.  I like Webhints / Shaky leaves as avenues to explore particularly Record hints and as mentioned, the Other Trees can provide links to other researchers.

 

I am now settled in Accepting Ancestry records that are clearly the right person even though there may be errors, as happens with normal evidential Records in our day-to-day research.

 

Cheers

Robert