How to properly combine improperly split families in a census record?
First, my apologies if this is answered somewhere, but I can't even figure out the proper search terms at the moment (especially given the recent interface changes). Now, my question: in the event that I find a family inappropriately split across multiple households in a census page, how do I combine them into one?
For instance: https://www.familysearch.org/ark:/61903/3:1:33S7-9R7Q-SW78
The Carrol family on rows 78-83 are split across three different households in the system, which makes it difficult to cross reference and make sure things are linked to the right people. Do I find one of the households, press Add Person, and just type it all in? I don't see an obvious tool for this.
Best Answer
Answers
-
Well huh, there's the Manage Indexes link I was unable to find before. Thank you very much, that's… exactly what I needed.
1 -
Forgive me for asking but - why should anyone spend time and effort correcting the six members of this family in the index entries? The aim is to get the profiles of these people correctly and they can be attached to the relevant lines on the page, without worrying about the households.
It seems to me that it's distinctly possible that the time and effort involved in adjusting the index entries outweighs any benefit in cross referencing. After all, let's think about it, many of us were working with microfilms at one time, which didn't have any index and we got through it quite happily, checking the cross references and "linking" / "attaching" by eye.
If someone is going to tell that it's important to establish why the household was split apart, I agree - but the suggested approach won't do this.
Having mixed up households may affect the "Unfinished Attachments" process but I don't know if the Carrol family were mixed up households, or just split too deeply. in any case, dismissing the "Unfinished Attachments" process does not affect finding any unfinished attachments.
As always, just because you can, doesn't mean that you should - particularly if the extra work outweighs any benefit to the finished profiles.
Please tell me if I'm missing something!
-1 -
@Adrian Bruce1 some people like their sources to be neat and tidy. To each his own…
1 -
Isn't this also about the ability of sources to verify family relationships, not just profiles? I know it's ridiculously long-winded to attach a source to a child/parent or couple relationship, and it's pretty much invisible once you've put it there (you can't even tag a C/P source), but it seems to me objectively a good thing to do it. (I have moaned before about the FT Details display making every transatlantic trip my grandfather took look more significant than his marriage and children.)
And tidying up the demarcation and indexing of one or more households on a census image can only have beneficial effects on the accuracy and retrievability of the other households shown there.
I would also hope that such corrections would feed the learning of the CAI demarcation model for that image collection.
2 -
In addition to Mandyshaw1's excellent explanation, in this particular case I've had a very persistent hint that conflates two different profiles:
https://www.familysearch.org/en/tree/person/details/G8JG-YPWhttps://www.familysearch.org/en/tree/person/details/G41S-F6D
Both are named Harold Carroll, both were born in Alabama in 1913, and both have separate families with separate census records. "Suggested records to review" page, however, is absolutely convinced that one Harold should be on both. It's much, MUCH easier to confidently say "no, that's not right" when the whole family is present on the source linker for each record at the same time. Atomized records make it much more difficult to be confident about the right move.2 -
@pauli133 - that's a good point, though my personal approach (bearing in mind @shecorwe's "to each his own") would be to look at the image rather than the source-linker. So many of those are hard work because they're based on different people than the way I'm approaching the issue. (The classic example of this is a marriage where the one set of parents that I want belong to the bride but the pair I get belong to the groom.) I can read the image faster and easier than the source-linker.
0 -
@MandyShaw1 - yes, that's two important points stated or implied there. Firstly, Data Quality checks assume the supremacy of the Source Index - which is not always true. Given that I've (presumably) completed the correct attachments of the Source Index, I prefer to just dismiss any erroneous DQ alerts rather than spend goodness knows how long rewriting the index while not touching the profiles of the people concerned.
As for "tidying up the demarcation and indexing of one or more households on a census image can only have beneficial effects on the accuracy and retrievability of the other households shown there", well, that depends on what the household demarcation issues were. If the Carrol family was split into (say) 3 households affecting none of the others, it doesn't make much difference. If the 3 (say) households included members of other households, then it might make sense.
As for your "hope that such corrections would feed the learning of the CAI demarcation model", well, yes. I suspect "hope" is the operative word in the absence of any clearer statements.
1
