The Quality Score
This web page;
is, for a first time user, completely useless. It doesn't explain the most basic things such as:
a. What is a "Tagged Source"
b. For William Wilson [G58Z-GQC] (my paternal 3g g.f.) as an example, only 3 "issues" are identified resulting in a "Low Quality" score. It would be nice if the "Data Quality Score" page explained the number of "issues" that go into calculating the "score".
c. Under "Conflict-free Data" is stated "The death happened before a marriage" but gives absolutely no details. Why can't the specific data be shown? If you have to limit the detail in the items making up the "Quality Score" why can't you give a web address that does provide all the detail underlying all the "issues" used in calculating the "Quality Score".
d. Under William Wilson's Quality Score are listed 4 items: 1. Data Completeness 2. Source Tagging 3. Source Consistency 4. Conflict-free Data. Are those the only categories? How many other categories are there? Can't they be listed in the Data Quality Score web page?
e. Why doesn't the "Data Quality Score" lay out what goes into the scoring. What constitutes "Low", I assume "Medium" and, again I am assuming "High". I realize scoring can probably get very complicated but it would be nice if a page that is supposed to "educate" a user about "quality scoring" be given some examples. For the William Wilson example, why is a missing burial counted as an "incomplete" when a death record has been included in "Sources"? Why can't your "Data Quality Score" enumerate the number of items that go into the "Data Completeness" score?
f. This "Ask a Question" page could use some clarification. What is the difference between the box immediately below "Question" and the following empty box which is bigger than the first? Should what I have been typing up to now have gone into the first empty box? If not, why can't you label the second empty box as "Detail" or "Elaboration". If the first box is NOT supposed to be the detailed question why can't it be labeled "Heading" or "Category" or "Type"? If the 2nd box is NOT supposed to be the actual question, what is supposed to go into it?
Respostas
-
I completely agree that FS should publish the algorithms (I do get that people may 'game' them, but those who are prepared to take care, and thus to work to optimise FT, need to understand exactly what they are looking at and why).
1 -
Your points are valid—the Data Quality score is a "heads up" to identify issues that might not be obvious
William Wilson https://www.familysearch.org/tree/person/G58Z-GQC marriage date to Anna Haseltine displays as 19 September 1819, but the standardized date [what the system sees for calculating] is 19 September 2018. Try editing the marriage date to see how it affects his quality score.
2 -
@scott wilson 1950 by chance did you review these links on the group homepage?
0 -
Just another user here who has no control over anything but going to make some comments anyway. The FamilySearch website is massive, has frequent updates, and often the help center articles are limited or are months to years behind.
f. Communities does have its issues. It seems to be a third party product with a FamilySearch skin on it and that third party doesn't seem to be very responsive to fixing things for FamilySearch. The first box is for the title and only holds a few dozen characters. The second box is for the actual question.
e. The entire Quality Score routine only recently came out of the FamilySearch Labs experiments page. It is limited in scope for both time and geography. A lot of development appears to still be going on. Maybe when it hits a more stable, developed phase so that what goes into the scoring quits changing, we'll get more details about what goes into the scoring. Right now, we users get to just experiment and see what makes a score change.
e. As far as the quality score is concerned, sources in Sources don't count as being present. They are only present if tagged to a specific data item. However, the information in all the sources, whether tagged or not, is compared to the data on a profile to look for data conflicts.
d. Those four categories are the only categories at present. There are no more. All four categories are always listed under Quality Score Details link even if no problems are found. As stated above, maybe we'll get fuller explanations in the information page when the routine quits changing.
c. Regarding the death before marriage issue, this is a duplicate of the Data Error also listed in the Research Help section:
If you click on the data error, you do get full information as to what the problem is as pointed out by @Miss Jessie
Since they can show it there, I would expect that the programmers should be able put the same information in the details for the Quality Score Details display of this. I would, however, expect this to take them six months to a year as they further develop the whole routine. (As a first time user, do you have any questions about Family Tree's magnificent dual entry of dates and places which incorporates both a Display date or place and a Standardized date or place? And the need to always make sure the the correct standard is linked to the displayed data? It is a very powerful system but can occasionally let inadvertent, somewhat hidden errors slip in as did here.)
b. It is not only the number of issues identified that result in a particular score but also the types of issues. For example, if the only quality issue identified is no burial information, this gives three issues but a High score. I assume this is because it recognized that burial dates are often just never found. In your example you also have just three issues, but that death before marriage date is a massive one and all by itself will always cause a low score.
a. Tagged Sources are sources that have been attached to a profile and so show on the Sources page that have also been tagged to a particular data item on the profile page and so show in the Data View windows which open when you click in the data field but not on the pencil edit icon.
The purpose of tagging is to let researchers find the supporting sources for any one piece of data without needing to search through the entire Sources page to find them.
For more about tagging, click on the ? icon in the top right of every page and type in a search:
I'll just mention that the Data View window is also an easy place to check whether the standards linked to the displayed data are correct or not:
You can also also check the standard by just hovering over the displayed data and waiting for the tool tip to appear:
This ability to do the same thing multiple different ways is very common in Family Tree.
6
