Citation confidence level

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

Citation confidence level

sophie1975
Hello,

I wonder if I used the confidence level for citations in the proper way
in the past. I always thought that it means the degree of reliability of
a source citation. This would be the correct meaning of the German Word
"Verlässligkeit" on the German client interface. But I just noticed that
on the narrative website it is translated with "Vertraulichkeit" which
means confidentiality, which is another meaning. In the documentation
the word confidence is used, that seems to stand for reliability.
So what is the real meaning of "confidence" reliability or
confidentiality?  Thanks in advance.

Nathalie


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Dave Scheipers
Hi Nathalie

I view the Confidence Level as... How confident am I that the
information I found at this source/citation applies to this person.

Yes, the citation is a birth record, with the date and location and it
was for a John Smith. Those are the facts of the citation and cannot
be disputed.. But how confident am I that this is my John Smith?

I cannot speak to the translation of the term. Maybe it is in error.
But then again, others may see the Confidence Level in other ways.

Dave

On Thu, Jan 3, 2019 at 7:30 PM sophie1975 <[hidden email]> wrote:

>
> Hello,
>
> I wonder if I used the confidence level for citations in the proper way
> in the past. I always thought that it means the degree of reliability of
> a source citation. This would be the correct meaning of the German Word
> "Verlässligkeit" on the German client interface. But I just noticed that
> on the narrative website it is translated with "Vertraulichkeit" which
> means confidentiality, which is another meaning. In the documentation
> the word confidence is used, that seems to stand for reliability.
> So what is the real meaning of "confidence" reliability or
> confidentiality?  Thanks in advance.
>
> Nathalie
>
>
> _______________________________________________
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Ron Johnson
In reply to this post by sophie1975
On 1/3/19 6:28 PM, sophie1975 wrote:

> Hello,
>
> I wonder if I used the confidence level for citations in the proper way
> in the past. I always thought that it means the degree of reliability of
> a source citation. This would be the correct meaning of the German Word
> "Verlässligkeit" on the German client interface. But I just noticed that
> on the narrative website it is translated with "Vertraulichkeit" which
> means confidentiality, which is another meaning. In the documentation
> the word confidence is used, that seems to stand for reliability.
> So what is the real meaning of "confidence" reliability or
> confidentiality?  Thanks in advance.

The English word "confidence" comes from the Latin word we get the word
"fidelity" (faithfulness or trust).  In this case, it's how much you trust
that the document is correct.

--
Angular momentum makes the world go 'round.


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Oliver Lehmann
In reply to this post by Dave Scheipers
Hi,

I use the confidence level more like to which grade the information in  
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence"  
level as the pastor could have made an error on copying the churchbook  
entry by hand
- other official documents reveal some data - this would be a "normal  
confidence" level to me
- my grandma tells me something she remembers or had written down  
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't  
attach it at all.

Best regards,
Oliver


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Philip Weiss
In GEDCOM, where a lot of these data items come from, confidence is how much confident the researcher is in the information as applied to the thing it documents.  GEDCOM has a separate entry of the information for every use of a record or document for every event, fact, name, or relationship.  This includes the confidence field.

So in that case I could have one reference to a death certificate to document a death, another entry to that same death certificate to document the birth, yet another to document the name, etc.  Each gets a different confidence level.  And that's exactly what I want.  I'm pretty confident the death certificate has the correct date of death.  Way less confident that it has the correct date of birth.

Gramps did something that was really useful in one way. There is only one entry of a citation.  It essentially means "document" or "record".  It means I only have to enter that death certificate once.  Even more importantly, if I need to edit it, I only need to edit it once and don't need to find every place that references the death certificate to update it.  For instance, I get a better, more legible copy of the death certificate and realize that where I thought it said 1943 it actually said 1948.  Re-using citation objects is AWESOME because of that.

However, right now a citation only has one confidence level field.  I can't set it to be high for the death and low at the same time for the birth.  At one point, I did use the field for my overall confidence in a record, but as I found more and more cases where that confidence wasn't the same, I just stopped.  Later, I went through and changed every citation confidence level to "Normal".

What I do now instead is use tags for events to indicate my overall confidence in the conclusion based on the totality of the evidence I have ("low", "medium", "high", and "gps").  Low, medium, and high are all educated guesses.  GPS is "very high" and means I'm satisfied I've achieved the Genealogical Proof Standard, particularly that I've got a written explanation of my reasoning.

As always, I'm not saying people have to do it this way.  Just hoping that it's useful to some folks.

Phil.

On Jan 3 2019, at 10:45 pm, Oliver Lehmann <[hidden email]> wrote:
Hi,

I use the confidence level more like to which grade the information in
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence"
level as the pastor could have made an error on copying the churchbook
entry by hand
- other official documents reveal some data - this would be a "normal
confidence" level to me
- my grandma tells me something she remembers or had written down
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't
attach it at all.

Best regards,
Oliver


_______________________________________________
Gramps-users mailing list
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

sophie1975
Thank you all for your answers. It seems that my first understanding and the way i used the field was correct (fortunately). For the moment I am perfectly ok with one field per citation even if several information are hold in it. Maybe once I will have to do some more granular tagging for myself.

In the meantime I would suggest to replace "Vertraulichkeit" by "Verlässlichkeit" on the narrative web site to avoid misunderstanding. The person to contact here is Mirko Leonhäuser if I remember well?

Am 04.01.2019 um 08:59 schrieb Philip Weiss:
In GEDCOM, where a lot of these data items come from, confidence is how much confident the researcher is in the information as applied to the thing it documents.  GEDCOM has a separate entry of the information for every use of a record or document for every event, fact, name, or relationship.  This includes the confidence field.

So in that case I could have one reference to a death certificate to document a death, another entry to that same death certificate to document the birth, yet another to document the name, etc.  Each gets a different confidence level.  And that's exactly what I want.  I'm pretty confident the death certificate has the correct date of death.  Way less confident that it has the correct date of birth.

Gramps did something that was really useful in one way. There is only one entry of a citation.  It essentially means "document" or "record".  It means I only have to enter that death certificate once.  Even more importantly, if I need to edit it, I only need to edit it once and don't need to find every place that references the death certificate to update it.  For instance, I get a better, more legible copy of the death certificate and realize that where I thought it said 1943 it actually said 1948.  Re-using citation objects is AWESOME because of that.

However, right now a citation only has one confidence level field.  I can't set it to be high for the death and low at the same time for the birth.  At one point, I did use the field for my overall confidence in a record, but as I found more and more cases where that confidence wasn't the same, I just stopped.  Later, I went through and changed every citation confidence level to "Normal".

What I do now instead is use tags for events to indicate my overall confidence in the conclusion based on the totality of the evidence I have ("low", "medium", "high", and "gps").  Low, medium, and high are all educated guesses.  GPS is "very high" and means I'm satisfied I've achieved the Genealogical Proof Standard, particularly that I've got a written explanation of my reasoning.

As always, I'm not saying people have to do it this way.  Just hoping that it's useful to some folks.

Phil.

On Jan 3 2019, at 10:45 pm, Oliver Lehmann [hidden email] wrote:
Hi,

I use the confidence level more like to which grade the information in
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence"
level as the pastor could have made an error on copying the churchbook
entry by hand
- other official documents reveal some data - this would be a "normal
confidence" level to me
- my grandma tells me something she remembers or had written down
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't
attach it at all.

Best regards,
Oliver


_______________________________________________
Gramps-users mailing list


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org



_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

ACProctor
In reply to this post by Philip Weiss

In traditional genealogical research (by which I simply mean 'not steered by GEDCOM'), this sort of information is part of the body of the written work, i.e. where inferences and conclusions are reached from the information in those sources. For instance, a common misconception is that cited sources deliver information, or so-called "facts", that you have used, whereas they're simply a record of the information sources that you've consulted. Hence, that information may be conflicting, or may be used to create an argument against it.

The citations themselves may have 'analytical comments' attached (e.g. layer1; layer2; ...; comments) but they're usually related to the nature and provenance of the source than the content within.

I'm not sure how this relates to the 'citation confidence' field, but when Gramps supports rich-text with mark-up then it would be capable of supporting the methods used by traditional genealogy.

Tony

On 04/01/2019 07:59, Philip Weiss wrote:
In GEDCOM, where a lot of these data items come from, confidence is how much confident the researcher is in the information as applied to the thing it documents.  GEDCOM has a separate entry of the information for every use of a record or document for every event, fact, name, or relationship.  This includes the confidence field.

So in that case I could have one reference to a death certificate to document a death, another entry to that same death certificate to document the birth, yet another to document the name, etc.  Each gets a different confidence level.  And that's exactly what I want.  I'm pretty confident the death certificate has the correct date of death.  Way less confident that it has the correct date of birth.

Gramps did something that was really useful in one way. There is only one entry of a citation.  It essentially means "document" or "record".  It means I only have to enter that death certificate once.  Even more importantly, if I need to edit it, I only need to edit it once and don't need to find every place that references the death certificate to update it.  For instance, I get a better, more legible copy of the death certificate and realize that where I thought it said 1943 it actually said 1948.  Re-using citation objects is AWESOME because of that.

However, right now a citation only has one confidence level field.  I can't set it to be high for the death and low at the same time for the birth.  At one point, I did use the field for my overall confidence in a record, but as I found more and more cases where that confidence wasn't the same, I just stopped.  Later, I went through and changed every citation confidence level to "Normal".

What I do now instead is use tags for events to indicate my overall confidence in the conclusion based on the totality of the evidence I have ("low", "medium", "high", and "gps").  Low, medium, and high are all educated guesses.  GPS is "very high" and means I'm satisfied I've achieved the Genealogical Proof Standard, particularly that I've got a written explanation of my reasoning.

As always, I'm not saying people have to do it this way.  Just hoping that it's useful to some folks.

Phil.

On Jan 3 2019, at 10:45 pm, Oliver Lehmann [hidden email] wrote:
Hi,

I use the confidence level more like to which grade the information in
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence"
level as the pastor could have made an error on copying the churchbook
entry by hand
- other official documents reveal some data - this would be a "normal
confidence" level to me
- my grandma tells me something she remembers or had written down
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't
attach it at all.

Best regards,
Oliver


_______________________________________________
Gramps-users mailing list


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

gramps-users
In reply to this post by sophie1975
Hi!

I read this thread and am already working on a fix.

Kind regards,
Christian




Am 04.01.19 um 09:33 schrieb sophie1975:

> Thank you all for your answers. It seems that my first understanding and
> the way i used the field was correct (fortunately). For the moment I am
> perfectly ok with one field per citation even if several information are
> hold in it. Maybe once I will have to do some more granular tagging for
> myself.
>
> In the meantime I would suggest to replace "Vertraulichkeit" by
> "Verlässlichkeit" on the narrative web site to avoid misunderstanding.
> The person to contact here is Mirko Leonhäuser if I remember well?
>


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

GRAMPS - User mailing list
In reply to this post by Oliver Lehmann
I'd like to see "Fallacy" as one of the levels of confidence.

Once a "Fact" is incorrectly attributed to a person, it will be cited each time another historian finds another reference to that knowledge. And each citation lends more credence 

If we disprove something and then discard the proof, we will probably have to disprove it again in the future.

As an example, an 18th century biographer wrote a passage about an Irish ancestor sibling immigrating to a county in western Pennsylvania after the American Revolution. The writer meant that he, after being mustered out of the Continental Army, finally settled in that county. But, because later researchers assumed "immigration" could only relate to movement from from outside the US rather than from different settlements of the Colonies, they "confirmed" the date his Ship arrived from Ireland in 1783. We've since confirmed that he enlisted in another county in eastern Pennsylvania at the beginning of hostilities AND he had a child born there before enlisting. We no longer know his year of Emigration, or how long he had been in the Colonies. They could have been born there and have been of Irish extraction.

But, because misinterpretation of the Imigration is 'known' from a contemporary source and the new information makes the Ireland Emigration 'unknown', we go through the exercise of disproving this Fallacy recursively. 

(The same thing happens with Julian date double conversions and Baptismal records that are assumed to be Birth dates & places. A 'pie in the sky' wish is to be able to set a "not equal to" in Gramps so that that data value will be refuted during data entry the NEXT time it pops up. )

So it would be nice to be able store the Source of a Fallacy datapoint.  Then, when the fallacy Source is cited, the conclusion can be more quickly discounted. But how do you do this without "fallacy" sources being added to Gramps reports... adding credence to the misinterpretation? 

After all, information that has been "Privacy" locked will still show up in reports when the data filter doesn't take that tag into account.  

On Fri, Jan 4, 2019 at 1:14, Oliver Lehmann
Hi,

I use the confidence level more like to which grade the information in 
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence" 
level as the pastor could have made an error on copying the churchbook 
entry by hand
- other official documents reveal some data - this would be a "normal 
confidence" level to me
- my grandma tells me something she remembers or had written down 
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't 
attach it at all.

Best regards,
Oliver



_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org

Untitled (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Ron Johnson

My first thought would be to make a custom event "Immigration Fallacy".

On 1/5/19 11:27 AM, Emyoulation--- via Gramps-users wrote:
I'd like to see "Fallacy" as one of the levels of confidence.

Once a "Fact" is incorrectly attributed to a person, it will be cited each time another historian finds another reference to that knowledge. And each citation lends more credence 

If we disprove something and then discard the proof, we will probably have to disprove it again in the future.

As an example, an 18th century biographer wrote a passage about an Irish ancestor sibling immigrating to a county in western Pennsylvania after the American Revolution. The writer meant that he, after being mustered out of the Continental Army, finally settled in that county. But, because later researchers assumed "immigration" could only relate to movement from from outside the US rather than from different settlements of the Colonies, they "confirmed" the date his Ship arrived from Ireland in 1783. We've since confirmed that he enlisted in another county in eastern Pennsylvania at the beginning of hostilities AND he had a child born there before enlisting. We no longer know his year of Emigration, or how long he had been in the Colonies. They could have been born there and have been of Irish extraction.

But, because misinterpretation of the Imigration is 'known' from a contemporary source and the new information makes the Ireland Emigration 'unknown', we go through the exercise of disproving this Fallacy recursively. 

(The same thing happens with Julian date double conversions and Baptismal records that are assumed to be Birth dates & places. A 'pie in the sky' wish is to be able to set a "not equal to" in Gramps so that that data value will be refuted during data entry the NEXT time it pops up. )

So it would be nice to be able store the Source of a Fallacy datapoint.  Then, when the fallacy Source is cited, the conclusion can be more quickly discounted. But how do you do this without "fallacy" sources being added to Gramps reports... adding credence to the misinterpretation? 

After all, information that has been "Privacy" locked will still show up in reports when the data filter doesn't take that tag into account.  

On Fri, Jan 4, 2019 at 1:14, Oliver Lehmann
Hi,

I use the confidence level more like to which grade the information in 
the document is correct.
For example:

- a Scan/Copy of the original churchbook gets a "very high confidence" level
- a baptism certificate made by the pastor gets a "high confidence" 
level as the pastor could have made an error on copying the churchbook 
entry by hand
- other official documents reveal some data - this would be a "normal 
confidence" level to me
- my grandma tells me something she remembers or had written down 
somewhere - "low confidence" level...

This is how I use this field.
When I'm not sure if the document is for the person at all I won't 
attach it at all.

Best regards,
Oliver


--
Angular momentum makes the world go 'round.


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

GRAMPS - User mailing list
That is certainly an option, but given that there are centuries of record keeping & research, there are errors everywhere. It would would result in a custom 'fallacy' counter event for every standard event. That's not sustainable.

On Sat, Jan 5, 2019 at 13:45, Ron Johnson
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org

Untitled Download Attachment
Untitled (268 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

Ron Johnson

The correct method would be to have add a field to each event which flags whether or not it really happened.  That would require a lot of programming changes (not only to the Event Editor window but all reports).

Adding Fallacy counterparts when required would be much simpler.

On 1/5/19 1:56 PM, [hidden email] wrote:
That is certainly an option, but given that there are centuries of record keeping & research, there are errors everywhere. It would would result in a custom 'fallacy' counter event for every standard event. That's not sustainable.



--
Angular momentum makes the world go 'round.


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Citation confidence level

GRAMPS - User mailing list
In reply to this post by GRAMPS - User mailing list
Come to think of it, an additional stronger negative confidence level might be warranted: "Falsification"

Lots of 18th-20th century people trying set themselves up as Society or Nobility fudged their pedigrees. (And dishonest genealogists looking for a Payday would bilk customers the same way.) Sorta like post-1800 rustic Inns with historic markers purporting "George Washington slept here." DNA is exposing some of the long accepted ones.

I'm reminded of the Samuel Clemens (Mark Twain) autobiography bon mot: "There are three kinds of lies: lies, damned lies, and statistics"

On Sat, Jan 5, 2019 at 13:59, Emyoulation--- via Gramps-users
That is certainly an option, but given that there are centuries of record keeping & research, there are errors everywhere. It would would result in a custom 'fallacy' counter event for every standard event. That's not sustainable.
________________________________________


_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org