Category Structure Proposal: Evidence lacking for locations

Question

Following a proposal by Ian Beacall to deal with birth and death events where there is no evidence for location, I suggested using categories for this purpose. Chris Whitten asked me to pursue the idea with a new proposal. The current post is the first one along these lines in the Proposing Category Structures process.

In brief, the proposal is for new categories, either (in variant 1)

• Category: No Evidence Available for Birth Location
• Category: No Evidence Available for Death Location

or (in variant 2)

• Category: Both Birth and Death Location Lack Evidence

The draft proposal is at Category structure proposal: Evidence lacking for locations. Please review it and answer or comment below, with a particular focus on:

1. The category names: Suggestions for improvement of these are welcome. The names need to meet the general category naming rules.
2. Structure: What should the parent category be: Maintenance Categories, or something else? Do the proposed categories duplicate or conflict with an already existing category structure?
3. Projects: Do the categories overlap or affect any project?
4. Data Doctor suggestions: Consider the relation of the categories in the proposal to possible future DBE suggestions as planned by Aleš Trtnik.
5. Choice: Please express your preference between variant 1 and variant 2 of the proposal.

After any modifications to the draft proposal arising out of discussion here, and selection of the more popular variant, I will proceed to the next step, Formal G2G Proposal, in a new post.

Answer 1

Option one appears to be preferable as it provides more precision, and I am assuming Alex's error reports will deal with each location field separately, so that a profile with a birth but not a death location will generate an error message.
 However for consistency should evidence be replaced by source? That would automatically link in to the source definitions on Wikitree. I can't see that evidence is any different to source.
 So something more along the lines of "No sources can be found for _______ location'

Comments

Thank you for these points, Gary.

Aleš's proposed suggestions would apply only to profiles with neither birth nor death location, and would indicate that both the birth and death location fields together are blank, without distinguishing between or separating them. In this sense variant 2 of the current proposal is quite compatible with the suggestions.

I am proposing the word "evidence" because I do not believe that "evidence" is the same as "source". Consider as a concrete example a hypothetical couple A and B for whom there is a marriage source in England. Say that they then had three children C, D and E over three years. There are sources placing the births of C and E at a location in England, but no birth source has been found for D. Nonetheless it would be reasonable, and better than a mere guess, to set the location field for the birth of child D to "England", with status "uncertain". This would be based not on a source, but on the indirect evidence provided by the records for the other family members A, B, C and E.

---

For suggestions it doesn't really matter. I can exclude one category or a list of categories or  even a category with all subcategories. Whatever is needed.

Answer 2

I agree Option 1 provides more precision.

Will something be set up so that we do not receive two warnings about missing birth and death locations as Error 574 for Find-A-Grave also address these?  Once as an Error 574 and once as the new one?

Comments

Good question, M. That would be up to Aleš to determine if and when he goes ahead with the new suggestions he proposed.

---

Suggestion 574 will probably not use these categories. It may provide a lead that a location is known. Suggestion can even pop up in the future, when someone enters a location on FG and a profile no longer needs to be in the category.

Answer 3

I'm a bit concerned about overlap with the "Needs Birth Record" and "Needs Birth" categories, that already exist. How about replacing "Location" with "Country" in the proposal?

Comments

The "No Evidence Available for Birth Location" category would not overlap with "Needs Birth Record", but be just about opposite to it. "Needs Birth Record" means that research needs to be done to look for a birth record, including location. "No Evidence Available for Birth Location" would mean that research has been done which establishes that neither a birth record with location, nor other evidence of birth location, is available.

I'm not sure what you mean by the "Needs Birth" category, Florian: as far as I know there isn't one.

As pointed out on the Uncertain Locations Help page, even a continent is better than no location, so "country" is not the right term. Instead, the term "location" is used in the proposal because the new categories are there as an explanation of why the corresponding location field is blank.

---

First: I meant categories like Category: France, Needs Birth, Category: Spain Project Needs Birth, but those indeed don't seem to be that many.
-----
Second: At least for German emigrants it often appears that the  "Research was done" by the immigrant descendants often does not compare to the possibilities the local researchers have, so I think there would be way too many profiles in those "lack evidence" categories. Also folks using the category picker might just take it as "yes, I don't have evidence for it at the moment" or "the source I have sucks, I don't consider it as evidence". When sorting Category: Germany, Needs Birth Record, I usually put the profiles with only a country or a big region into Category: Germany Family Brick Walls, which in my opinion for the moment covers the same situation as the "lack evidence" category. I doubt there would be people doing the same sorting for the lack evidence category. Therefore it will just clutter up, like some of the brick wall categories do already. Where's the difference? Where's the need?

---

If a profile is in Category: France Needs Birth, then it is already known with a reasonable degree of confidence that the birth occurred in France. The birth location should not be blank: it should be "France", or a region of France (as is already the case for the overwhelming majority of profiles in that category). There is some degree of evidence for birth location. The birth location field is not blank; none of Aleš's new 14x suggestions can appear; and Category: No Evidence Available for Birth Location should not be applied. So there will be no overlap.

The pages for the proposed categories would contain instructions that they should only be used in rare circumstances, where careful research has been done. This can be supplemented with advice to document the research done within the profile, usually in Research Notes. If people apply the proposed categories contrary to the instructions, then just as with any other mistake on WikiTree, someone else can reverse it. If the same person persistently breaches the guidelines, a Mentor Intervention Request may be appropriate. Any rule can be broken, but not with impunity, so this is not an argument against having a rule.

The need is to provide profile managers with a way of forestalling the proposed new 14x suggestions in cases where a location genuinely cannot be entered, instead of having to undertake the labour of watching out for them and then retroactively marking them false.

---

Jim, the project maintenance categories have one for Needs Birth and another for Needs Death record.  Both the Acadians Project and Quebec project have them so I would guess there isn't just one but many, if other projects adopted the same set of categories that were proposed at the time.

I do like your differentiation of Evidence from Source.  I have recently begun using evidence for birth locations based on sourced family residences that bracket the year of birth.

Hopefully people won't mark Lack of Evidence as a substitute for Lack of Trying to find Evidence . . .

Answer 4

Neither Option 1 / Option 2 Yet: need more information.

Comments

What are the implications of variants 1 and 2 appearing to be inconsistent with the Category Names guideline below?

• If a category could contain millions of people create a narrower subcategory. Ideally, bottom-level categories should be much smaller, with no more than a few thousand people in them. This isn't always possible, e.g. with towns and cities, but we should aim to create narrow categories when we can.

Will the category pages be useful / responsive with the expected number of profiles?

---

The stated (below) uncertainty in purpose of the new categories makes review more complex. If the purpose is to forestall DD suggestions, then the category name could be very specific to DD suggestions, perhaps a new subcategory of categories just for that DD purpose. If the purpose is to address uncertainty in the birth and death locations, then the wording of the category descriptions might well influence the preference between variants.

Category size 

If the new categories are mainly used for the purpose of forestalling the planned 14x Data Doctor suggestions, their sizes will be quite limited, because they will only be needed in cases where both birth and death location lack any evidence. If they are used more generally, they could become large. If this is considered a risk, Proposal Variant 2 below could avoid it.

---

Would the new categories be used for cases where a birth or death location is listed, but there is no evidence for the location (including no research note with a circumstantial case for the location)?

Perhaps my question is misworded. Proposal variant 1 says it is for situation where the location would be a guess, but does that then mean that guesses should be removed and the proposed category added?

---

Thanks Ken. These are good questions.

I'll address your third and last comment now, and leave the rest till morning my time: it's late here.

1. The new categories are only intended for use where the corresponding location field is blank. Perhaps I can add wording to the space page to make this clearer.

2. A strict interpretation of WikiTree guidelines means that locations should not be guessed. As a corollary, guessed locations should be removed, and then the proposed category could be added provided the required research had been done.

See the Help page section Almost Any Location is Better than No Location, which says "The only exception would be if entering any location at all would be a guess."

---

Re Ken's second comment, particularly "If the purpose is to forestall DD suggestions, then the category name could be very specific to DD suggestions, perhaps a new subcategory of categories just for that DD purpose":

This is an interesting idea. There are other cases where it would be useful to have a way of forestalling Data Doctor suggestions which the profile manager knows will appear but be false. A general mechanism, maybe something like "Category: Forbid DBE 123", could be considered. However, this is beyond the scope of the current topic, and I'm afraid that discussing it in detail here would derail the present proposal. Let's leave it for another day. If anyone wants to pursue it, please open a separate G2G question.

---

Re Ken's first comment on category size:

I do not think the single category in Variant 2 would become very large. Profiles should only be categorised there if neither birth location nor death location can be found (already not very common), and research has been done to demonstrate this in both cases (so the situation will be rarer still).

For this reason I personally favour Variant 2. Though I appreciate the arguments of people on the thread who prefer Variant 1 because of its extra precision, I think this may need to be sacrificed to meet the category-size criterion Ken has emphasised.

An option might be to start with Variant 1, and monitor the situation. If the two Variant 1 categories became large, perhaps EditBot would be able to move the profiles which are in both of them to a single category like Variant 2, and remove profiles from either category where they were only in one. That way we could resolve the size problem by converting from Variant 1 to Variant 2 at a later date.

---

Most people who've commented favour Variant 1, with two separate categories for birth and death. Concern remains, as here from Ken, that such categories could become very large. However, Aleš has indicated that in that case it will be possible to change retroactively from Variant 1 to Variant 2 with a single much smaller category, either manually with assistance from WikiTree Plus or through an enhancement to EditBot. Therefore it seems safe to proceed with the more popular Variant 1 at this stage.

Answer 5

At presently worded:

Variant # 1 is my choice, it makes more sense as it clearly states that there is no evidence available which is why no location has been entered.

Category: No Evidence Available for Birth Location

Category: No Evidence Available for Death Location

Variant #2  Both Birth and Death Location Lack Evidence the wording seems to imply that a Birth and a Death location have been entered and have no documentation. 

I know that was not the intent of Aleš original idea. " I decided it is finally time to flag profiles with no Birth or Death location entered as a suggestion" 

https://www.wikitree.com/g2g/1695136/new-suggestion-on-the-horizon

Variant #2 could be edited to state No evidence available for Birth and Death locations 

If that change was made to the description of Variant # 2, it would be preferable as it fits the purpose of Aleš original idea. 

I think it is important that a person who may work on the new categories realises that significant research has already been done to find a birth or death location, and that no evidence has been found either from a source or that could be inferred from other family members. 

There will always be profiles that cannot be further improved as the sources necessary are not currently available or do not exist at all.

Comments

Thank you M for these points, and in particular for the suggestion of renaming the Variant 2 category. We could indeed consider "No Evidence Available for Birth and Death Locations" (title case is required), provided it seems clear. I had thought of "No Evidence Available for Either Birth or Death Location", but it was too much of a mouthful. Your idea is better.

---

I like variant 1 as well to be used in profiles.

It doesn't seem right to use #2 in profiles as it will require too much changing when one of them gets completed but the other one isn't.  And where would it go if it's for both.  However, I DO like it for Ales to use as a suggestion.

Answer 6

Variant 1 seems better to me.  And I think Maintenance Categories for the parent sounds logical.

Answer 7

Neither.  

To my mind, if there is absolutely no evidence of where a person was born, such as is often the case with children whose parents migrated from one country to another, say from Germany to the USA, where the child could have been born in the country of origin, in a country between Germany and the port of departure, at sea, or after arrival in the US at the port of entry, and there's no record showing where, then it's a bio item.  I certainly don't see that a category will be useful for this.  

Same principle applies to place of death.

The DD error for this lack of location could be amended so that, after due diligence, when it has been determined that we don't know and never will know where, the location could have Unknown entered into it. Right now Unknown is not allowed in these fields.

There are oodles of profiles that I have seen created long ago that have location boxes blank.  The majority of them on research wind up having at least a country / region possible to enter.  It's a bio item, or Researh notes item.

I certainly don't think an additional category line will be useful for these.

Comments

>  if there is absolutely no evidence of where a person was born, such as is often the case with children whose parents migrated from one country to another, say from Germany to the USA, where the child could have been born in the country of origin, in a country between Germany and the port of departure, at sea, or after arrival in the US at the port of entry, and there's no record showing where, then it's a bio item.

But that is the point of this category.  This is for relatively rare cases like this.  There are new suggestions coming up that will flag blank locations as 'errors' as they should at least have a continent.  This proposal/category is for cases where we don't know (cannot know?) what the continent was.  The (main?) point of the category is to prevent these profiles from getting DBEs about the empty locations.  I'm not sure how you can object to this idea.  

---

My point is that a category is not the way to handle this.  Categories are meant to group profiles.  Maintenance categories are used to draw attention to the fact that something needs more data / sources, whatever.  Once the work has been done, the category is removed.  I see no sense in grouping profiles that are disrelated totally and for which there won't be an answer.

A better way to my mind would be the use of a Keyword in location fields, such as has been discussed in relation to Unknown as LNAB.  DBE recognizes keywords and can act accordingly.

---

I see.  And, yes, I agree with your point about categories in general. They're usually used to group things for useful purposes.

The keyword idea was actually Jim's first idea , which got derailed when I thought a better idea would be to use the data status buttons (because that's what this is about). My proposal was rejected by the Team, so we're here... The idea of using 'Unknown' was rejected . So if a keyword system is the best way, it just needs to be a different keyword.

For what it's worth, I think the category idea is fine as it's a maintenance category, which is a bit different from categories like location categories.  Also, it takes slightly more effort to add this category than to click a button or type a key word, which is a good thing in this case.

---

I've no problem with it being a different Keyword than Unknown, I just threw that one out there since it's familiar to everybody.

And it's just as easy to use category picker and click this proposed category, than to type in a keyword.

Setting this up as a maintenance category is likely to frustrate people who work on maintenance categories deliberately.  It would be a no-win situation.

---

As Ian points out, the idea of using a keyword or key phrase has already been proposed and failed. To achieve a solution, we need to move forwards, not go round in circles. Using categories to forestall suggestions has precedents, and Chris has expressed interest in the category approach for this case.

I see Danielle's point against Maintenance Categories for the parent. Perhaps a less controversial parent might be Category: Family, by analogy with its already existing subcategory Family Brickwalls. Absence of evidence about birth is a kind of brick wall. Absence of evidence about death is similar: the historical disappearance of a person without any known death information is a sad event which affects families.

---

If I understand correctly, this is all stemming from DBE reports that outline the lack of a location in the boxes.

When there is no data whatsoever to fill in the blank, one can't help but leave it empty, although I have seen authors filling in the blank by using last known place of residence (I know Jetté does this).  But such remain unproven, and often the person went someplace else in the meantime, as evidenced by notarial acts or other documents.  So it's a catch-22 situation: either fill in erroneous data or leave it blank and generate an error.

One thing I have noticed is that if you mark something as false error, for some reason it will get checked again at a later date, necessitating repetition of false error marking (unless this has been fixed since I last saw it happen).  So this category would appear to be a stop-gap measure so the program quits repeating its DBE reports on the same profiles.

If this is the case, then I would suggest it be named ''Category: Unknowable Location'' (birth / death can be subs), and be under the Brick Walls as parent.  That way, there won't be DBEs generated, and they won't be re-looked at constantly by those working on maintenance categories.

---

Hi Danielle. If there was an issue such as you describe with false suggestions, I believe it has gone away. These days, if a suggestion is marked false, it is hidden permanently, and no longer appears by default in the list of suggestions for a profile or profile manager. It will only become visible if the "Show Hidden suggestions" option is selected.

Category names with "Unknowable Location" could be considered, but at the moment I favour names with the words "No Evidence Available" to try to emphasise that before the categories are used, research must be done to seek evidence, and fail to find it.

I'm not sure the category Family Brick Walls is suitable as a parent, because at present it has a structure of subcategories organised by geography and surname. Extra special-purpose categories might not fit well. There is no "Brick Walls" category as such. That is why I suggested Category: Family as a possible alternative parent for the new categories in this proposal.

---

Maybe Category:Family Mysteries would better fit as parent?

The reason I suggested Unknowable Locations is that I find ''No Evidence Avaiiable for Birth Location'' rather long.

Answer 8

Yes I like option one best. You may find for a birth but not for the death, or vice versus. Thank you for this creation.