Data Completness Should Vary Based on Time Periods and Geography
First, I like the idea behind the Profile Quality Score, but the Data Completeness should vary based on time-period and geography.
Experienced genealogists know that in most cases as we go further back in time many record types are unavailable. In my case I research a great deal in Colonial Virginia. The idea that I should have a birth record or a death or burial record for persons living pre-1776, or pre-1850 for that matter, just doesn't make sense. There are however records that can prove a death or birth, but often only by year. For example, the will or estate inventory of a person who lived during this time-period is proof of their death, but usually there is no date of death included with these records, hence the date may be listed as before 5 February 1735 if a will was presented in court on that date. The current Profile Quality Score doesn't seem to recognize documentary evidence of this type, but for successful research during this time-period, in this place, probate records should be considered of very high value.
To expect a burial record pre-1850 in the United States is, in general, a totally unrealistic expectation. Most people buried pre-1850 in the U.S., do not have surviving grave markers, let alone those buried in Europe.
Would it be possible to adjust the Data Completeness information to account for the reality of existing records before 1850 in the U.S.?
Comments
-
While I'm sympathetic to the idea that many of these things in some areas and eras are "a totally unrealistic expectation", I feel that there is a danger of confusing two different but related things.
- Data Completion - If a burial record is missing for pre-1850 US profiles, then it is missing - the data is incomplete.
- Researcher's Effort - If someone can't find such a burial record, then they've almost certainly have done the best that they can and shouldn't be criticised for it.
I'm basically of the opinion that the Profile Quality Score is about what's on the profile, i.e. it's about Data Completion. (Feel free to dispute if you think it's about Researcher's Effort. Or even both.) As such we shouldn't imagine that it's our work that's being assessed, and we should be relaxed about a profile being incomplete if it genuinely is.
While the idea that the system should automatically apply completeness criteria, is initially attractive, I really fear that it's an impossible task for the engineers / FS staff to carry out since there are all sorts of subtleties. For instance, in Ireland, civil registration was introduced in 1864, hence the likelihood of finding marriage records pre 1864 varies incredibly from no-chance to maybe. Except, if the couple married in the Church of Ireland after 1845, then the marriage was registered.
That's getting complicated so the best person to assess this is the researcher. Perhaps they are the people who should dismiss "marriage missing" warnings, although I can't decide whether dismissing such messages is appropriate or not given my position above.
0 -
My issue isn't so much that a specific record isn't available as it seems to be the use of records outside of FamilySearch's indexed records to support an event. An example is a will or estate inventory as proof of death. This type of record should be treated as equal to an indexed record found on FamilySearch in the Profile Quality Score and Data Completeness. Probate records are often more complete and accurate than death records for proof of a death, and they were well kept and available, but because they aren't an indexed record when they are attached as proof of death they don't seem to be recognized by the system as being of equal value to an indexed death record.
Couldn't the Profile Quality Score be adjusted for pre-1850 U.S. ancestors? I research in England, Germany and the U.S. In Germany it's expected and very possible to find birth, marriage and death records for most locations going back into the 1600s. In England it's realistic to find the same back to about 1800 or maybe even 1700. But, other than in New England, in the U.S., it's unrealistic to expect birth and death records prior to 1850. The Profile Quality Score makes us feel good when we see a High Score, but it's frustrating to see a Medium Score when there are multiple records attached and tagged, but most are unindexed and are not typical birth, death and marriage records.
Isn't the Score supposed to give the user a sense of whether the research done on an ancestor has been thorough? That sources are attached and tagged? That events are supported by sources?
It's challenging enough to research these colonial U.S. ancestors, the system should recognize quality research, sources, etc. For example, I have a church birth record for an ancestor born in 1752 attached to my ancestor (which is unusual in itself) by the quality score for the birth is "Medium." What does it take to get a high score for a birth? Multiple records? Not going to happen for early U.S. ancestors.
If the quality score can't reflect that thorough research has been done, sources have been found and attached and tagged, and relationships proved, maybe the score isn't of much value.
0 -
@Kimball G. Carter - "… the use of records outside of FamilySearch's indexed records to support an event. …"
That is an excellent point. I was welcoming of the High / Medium / Low ratings against Vital Events but I have said (as far as I remember) that we need to see explanatory text (like we do for the profile as a whole) because, yes, we've seen instances where we can't understand why one event is only "Medium" when another, apparently similar, gets "High".
At the same time, I'm seriously concerned over how the software would be able to recognise what data is being extracted from (say) a non-FS parish record. Logically, one would need to both tag the requisite event and supply the date (say) so that the quality scoring could know whether the date is a precise, single day, or a rough year. That's extra work - maybe if I want to persuade people of the quality of my research, then I'd complete the extra data. I suspect many wouldn't - and maybe that doesn't matter - if they're not interested to fill in the extra data, maybe they're not that interested in the score.
Actually, this might not just apply to records from outside FS - I've a feeling that I've seen FS indexed records that haven't indexed data items that I'd like to use. But whether such non-indexed data would be under the scope of Data Quality, I don't know.
As far adjusting the dates when Data Quality Scoring is applied, that would seem sensible if rating US profiles pre-1850 as Low or Medium were judged demoralising. The alternative view is that such profiles are Low or Medium but that's inevitable and doesn't imply the research has been inadequate.
0