Jump to content


Photo

Ancestors with Weak Evidence

weak evidence

  • Please log in to reply
15 replies to this topic

#1 cj1260

cj1260

    Member

  • Members
  • PipPip
  • 26 posts

Posted 28 July 2018 - 08:24 AM

I'm curious to get people's experiences/opinions on how they handle possible ancestors where the evidence is weak to fully support their relationship.

Up until now, I have only entered people into RM where I had a relatively high confidence of their relationship to the family based on strong evidence (i.e. government documents, written histories, etc.). Now that I'm getting back to the late 1700s and early 1800s with the family roots in Ireland, the evidence is getting sparser.

For example, I recently searched for a baptismal record for a known relative in Northern Ireland. I found a record that matches her name, occurs in a decade that would make sense, but it is in a town about 22 miles from where many of my family members came from. The baptismal record indicates her parents' names.

Since this is the only baptismal record that matches her name in that area and in that era, there is a possibility it is the correct one. However, it is only a possibility. Nothing about it proves that it is the correct record for my relative.

So my question is this: Do I enter the parents into RM off of this possibly correct baptismal record and just heavily annotate their names in regards to the flimsy evidence it is based on, or do I set the information aside in a folder until I can find further evidence to corroborate the names on the record?

Would like to hear how people handle this type of situation.

Thanks.



#2 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3487 posts

Posted 28 July 2018 - 12:09 PM

I wonder about the same thing from time to time, and I often wish that RM might make it easier to have two databases that were coordinated together somehow or other - one that was a working database including speculations and one that had strong evidence all around. I think that "proven" may be too strong a word to use in genealogy most of the time, but evidence can be very strong on the one hand, and it can also have varying degrees of weakness. I don't think that even such an esteemed authority as the Genealogical Proof Standard really gets at the messiness of a lot of genealogical data. It seems to me to have an underlying and unspoken assumption that you can always get to some sort of reasonable conclusion despite weak evidence. I don't think you can in many cases.

 

I tend to put people with weak evidence into my database with a note explaining all my disclaimers. But when I share my RM data with other trusted researchers, my "weak evidence" people can easily start showing up on the Internet as if there were no doubts about the evidence. It's a difficult problem.

 

Jerry



#3 Trebor22

Trebor22

    Advanced Member

  • Members
  • PipPipPip
  • 178 posts

Posted 28 July 2018 - 12:33 PM

I generally put 'hot leads' in the last 'proven' person(s) notes so in your example I would detail the baptism I had found in the childs notes but not add the parents in to my database and I would keep the person on my 'to do list' and look for more clues. I also tend to share 'to do's' with cousins also researching the same branch - many heads better than one,  but it is very much a personal choice and I would also welcome suggestions!



#4 Don Newcomb

Don Newcomb

    Advanced Member

  • Members
  • PipPipPip
  • 1042 posts

Posted 29 July 2018 - 07:24 AM

This goes to a feature I've requested frequently: The ability to assign "strength" values to the connections between people. In your case, on a scale of 1 to 10, your speculative connection might get 0.5 value. Whereas a line that has been well documented from original sources and confirmed by DNA, might get a 9. ITMT, I use color coding to indicate records with issues. RED is my chosen color for those people. 



#5 cj1260

cj1260

    Member

  • Members
  • PipPip
  • 26 posts

Posted 29 July 2018 - 08:13 AM

Jerry, Trebor and Don thank you all for the comments.

Jerry, you touched upon one of my concerns in regards to entering the data. I too have witnessed how guesses become facts once they get disbursed to other people.

I do like the idea of color coding. I think this might work well for my way of doing things. The color coding function is probably something I should use more often in RM.

Thanks again.



#6 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3387 posts

Posted 30 July 2018 - 06:16 AM

Would like to hear how people handle this type of situation.

Thanks.

 

There is no easy answer to record quality becoming less detailed and more tenuous as we go back through time although this is where I stress and refer to The DNA of Places and Locations. If your ancestor and my ancestor drank from the same well on the same day then it's a safe assumption they at least knew each other, so where was the well?

 

About 90% of my research is in Ireland and I live there so I understand the challenges, however Ireland has over 65,000 distinct land divisions called Townlands many of which are quite small and you will generally find these refered to in Parish records on baptisms and marriages. Whereas in England the reference to a person might be "John of Syderstone" where Syderstone is a Parish, in Ireland it gererally would be "John of Derrycaw" where Derrycaw is one of 31 Townlands within the Parish of Tartaraghan which itself is only 15.5 sq miles so a much more accurate geographic indicator. This combined with best approximated Dates and Alternate Name variations is as good as it gets.

 

I work slightly differently from you and I am lucky not to be working the Smith name  :huh:  I work 8 main family names and record all mentions within a targeted Parish or location into Rootsmagic. I always liken this to doing a jig saw puzzle, whilst one person might rummage around in the box looking for edge pieces I prefer to empty the whole thing out on the floor and get everything face up for examination. I have very few custom facts but one I do make good use of, especially on older records, is called "Research Notes" with Description field enabled, this used to be called "Known Fact" but it's simply valuable indicator information. It might contain a Place and Place Detail and the description field containing "mentioned on baptism of Jane Doe", I always approximate a Birth date so these things eventually start to fall into place.

 

How Places should be recorded is an endless discussion on these forums, I always maintain the geocoding should always transcend and text name whether it be combined Place Details and Place or split components and that Rootsmagic, and other software, should make better use of this simple mathematical proximity in suggesting matches. Whilst outside of any standard and not really recognizable by online services I decided to build my Places with more emphasis on the Parish name and the Irish Townlands being Place Details within the Parish, this has worked very well for me in my research. Problems abound from the endessly varying user Place notation styles and how administrative divisions change over time compounded by the personal preference of how users want Places to appear in reports, I can see solutions to those problems and preferences but it will take a long time to overcome them and convince users of a better way.

 

For your Irish records I would advise concentrating on the Townland and these have remained the same for several hundred years and are still noted by the land registry office, some Parishes may have merged, Protestant and Roman Catholic Parishes may differ slightly but the Townland has remained a reliable land division. You will find a good reference to Townlands over at townlands dot ie, the wildcard search is not great so where you cannot discern the Townland back out to Parish, Barony or whatever and read through the list of Townlands as spelling can be difficult.

 

The old image below was produced when I completed Stevenson research in the Parish of Tartaraghan from all available sources, ignore the red circle which was to indicate something at that time, the image below that shows a typical Church baptismal record showing the townland where the family lived and fortunately containing both parents names, which is often not the case.

 

mapping-parish.png

 

stevenson-derrycaw.png


“Your most unhappy customers are your greatest source of learning.” -Bill Gates

It's now time for discretion, trust, patience and support

 

User of Rootsmagic 7.5.9, Family Historian 6.2.7, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#7 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 8375 posts

Posted 30 July 2018 - 03:07 PM

This is how I handle it.

 

1. Research log with comments

 

2. Transcribe the document into the Source's Research Notes. In the comments section below I will note my conclusion about the record. With the Research Notes report I can see review the evidence I have and why I made that conclusion. 


Renee
RootsMagic

#8 cj1260

cj1260

    Member

  • Members
  • PipPip
  • 26 posts

Posted 31 July 2018 - 03:15 PM

Vyger, that's an interesting approach that you have and not one that I contemplated previously, but I can see how it may be of value to my Irish research.

Also, I much appreciate the response and the education. I'm still working on understanding the different land divisions in Ireland and your post certainly helped clear it up for me. I'm going to start paying closer attention to how I code these in Place/Place Details.

Thanks.



#9 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3487 posts

Posted 31 July 2018 - 06:48 PM

 

About 90% of my research is in Ireland and I live there so I understand the challenges, however Ireland has over 65,000 distinct land divisions called Townlands many of which are quite small and you will generally find these refered to in Parish records on baptisms and marriages. Whereas in England the reference to a person might be "John of Syderstone" where Syderstone is a Parish, in Ireland it gererally would be "John of Derrycaw" where Derrycaw is one of 31 Townlands within the Parish of Tartaraghan which itself is only 15.5 sq miles so a much more accurate geographic indicator. This combined with best approximated Dates and Alternate Name variations is as good as it gets.

 

I very much enjoyed reading this article. I didn't realize that there was anywhere in the world that place names had been stable over such a long period of time and where each place name represented such a relatively small area.  That in turn helped me understand more clearly the idea of using a single GPS coordinate as a surrogate for place names over time. But I don't think that idea would work at all for many of the place names that I research most heavily.

 

I research most heavily in Tennessee, Virginia, and North Carolina in that order. Most of my ancestors had moved to what is now Tennessee before Tennessee became a state in 1796. And in the 1700's, most of them were located either in Virginia or in North Carolina. During that time frame, there were radical changes in place names and the area which the place names represent. Indeed, prior to 1790, all of what is now Tennessee was part of North Carolina. The first county in what is now Tennessee was Washington County, and the original Washington County encompassed the entirety of what is now Tennessee. The original counties were few and huge and they gradually became broken up into more and smaller counties. A simple example of this issue in researching my Cross family is that Micajah Cross purchased land in Anderson County, Tennessee in 1805 and William Cross purchased land in Anderson County, Tennessee in 1817. But William's land from 1817 was in what is modern day Anderson County and Micajah's land from 1805 was in what is modern day Scott County, Tennessee - four or five counties away and over a hundred miles away. So a single pin does not work to represent North Carolina or Tennessee, or to represent Anderson County, Tennessee.

 

There are some wonderful animated county formation maps online which illustrate these issues very well. And it's even worse than it sounds because a county often was not operational for quite a few years after it was formed. Deeds and wills and marriages could often still be recorded in the previous county for several years until the new county actually had a courthouse and a government. So I really feel like I need to stick to place names as they were at the time of the event, and I need to manually create maps and place pins in the maps where the events really happened (or equivalently, to find manually the proper GPS coordinates). Therefore, I find RM's mapping tools to be of virtually no benefit to my research, nor do I find RM's automated geo-coding to be of much benefit to my research.

 

Jerry



#10 zhangrau

zhangrau

    Advanced Member

  • Members
  • PipPipPip
  • 1490 posts

Posted 31 July 2018 - 07:00 PM

Just to chime in - - -

 

When I have found people who might belong in a branch of my tree, but I'm not yet convinced of the certainty...

 

I add the person and associated data into my database, but I do not link them to the tree's likely branch. The linking awaits more definite data. But I consider that having the person in my database, as a possibility, encourages me to continue seeking clarifying and corroborating data.



#11 Trebor22

Trebor22

    Advanced Member

  • Members
  • PipPipPip
  • 178 posts

Posted 01 August 2018 - 04:13 AM

Just to chime in - - -

 

When I have found people who might belong in a branch of my tree, but I'm not yet convinced of the certainty...

 

I add the person and associated data into my database, but I do not link them to the tree's likely branch. The linking awaits more definite data. But I consider that having the person in my database, as a possibility, encourages me to continue seeking clarifying and corroborating data.

I guess this is basically what I was trying to achieve adding the info into a persons notes, my thinking is that the info is 'with the person' so easy to find but perhaps doing both might be worthwhile - I shall ponder :-)



#12 Don Newcomb

Don Newcomb

    Advanced Member

  • Members
  • PipPipPip
  • 1042 posts

Posted 02 August 2018 - 06:49 AM

Just to chime in - - -

 

When I have found people who might belong in a branch of my tree, but I'm not yet convinced of the certainty...

 

I add the person and associated data into my database, but I do not link them to the tree's likely branch. The linking awaits more definite data. But I consider that having the person in my database, as a possibility, encourages me to continue seeking clarifying and corroborating data.

 

I would always link a new person to something. In some cases I might have a conceptual ancestor "person" like "Possible Smith Family" to whom I'd attach potential Smiths. I live in dread of someone becoming unlinked from his tree and frequently check for random "floaters". Intentionally adding unlinked people to my database would be like throwing extra jigsaw puzzle parts I found on the floor into the box.  



#13 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3387 posts

Posted 02 August 2018 - 07:50 AM

Just to chime in - - -

 

When I have found people who might belong in a branch of my tree, but I'm not yet convinced of the certainty...

 

I add the person and associated data into my database, but I do not link them to the tree's likely branch. The linking awaits more definite data. But I consider that having the person in my database, as a possibility, encourages me to continue seeking clarifying and corroborating data.

 

I agree here and just like the pieces of the jig saw it's all about associations, except we are not dealing with bits of sky and buildings.

 

I do use Shared events as I find these useful where a family link is unproven. Recent example is I found a 1800's death of a hot contender in my wifes line, and the death was registered twice with a different informant on each registry. Now at that point the surnames of the informants meant nothing to me, and still don't, but both Place Details given by each informant were already in my database so there is some sort of geo association. One was male and one was female, the female could well have been a daughter or grand daughter using her married name where in UK & Ireland any hints to the birth name are generally dropped on marriage.

 

What I decided to do was enter both informants into my database with the address information they gave and Share the death fact of the person in question with these two individuals therefore providing a non family link for now.

 

It's a shame the "just type a name" is not searchable after entry but I think it's importand that these individuals show up in the index where there may become a more obvious family link in the future. I should say I also note any such unconnected individual the reason and association in the description field of a custom fact I have.

 

As regards floaters I maintain a number of groups by surname where the number of parents = 0 for further research.


“Your most unhappy customers are your greatest source of learning.” -Bill Gates

It's now time for discretion, trust, patience and support

 

User of Rootsmagic 7.5.9, Family Historian 6.2.7, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#14 zhangrau

zhangrau

    Advanced Member

  • Members
  • PipPipPip
  • 1490 posts

Posted 02 August 2018 - 05:58 PM

I've grown accustomed to thinking of SQL as a robust, not-error-prone, database. My main database currently has 423,781 people and 17,383 trees (the vast majority having 1 to 50 people). I do also use "suppositional" ancestors (such as {Ancestor] Smith) to collect together some of the unlinked records, but I've only been doing that for a year or so - and that's why I have so many unlinked trees.



#15 Rick Landrum

Rick Landrum

    Advanced Member

  • Members
  • PipPipPip
  • 347 posts

Posted 04 August 2018 - 08:39 AM

Posted 28 July 2018 - 09:24 AM

I'm curious to get people's experiences/opinions on how they handle possible ancestors where the evidence is weak to fully support their relationship.

 

I usually post "Not Confirmed" in the suffix field of the persons name. Like several others have replied, I then add notes in either the individual, or the source research, notes explaining why I posted them as "not confirmed". I go ahead and add them to my tree and set up a to-do task regarding any additional research needed.

 

I find that "flagging" the person in this way makes it easy to search and find persons that are not confirmed. I use color coding for other purposes, but a color could also be assigned to these persons making them stand out when reviewing your data base. You can also search by colors.

 

Hope this helps

Rick


RickL


#16 cj1260

cj1260

    Member

  • Members
  • PipPip
  • 26 posts

Posted 10 August 2018 - 10:50 AM

I've been on vacation for the past week or so, so when I returned I was happy to see that this conversation has carried on.

 

Thank you all for your responses. I've learned a lot how about how I can approach these types of possible ancestors and you all have really given me a lot to think about.

 

Much appreciated.