* Question re: usage of Census, Residence and Occupation

[Gary_G]

Teresa;

The discussion had progressed well beyond the point of my asking why a census was entered, if it couldn't be trusted.

I think that the more recent part of the thread was focussed on the analysis of what one records; something I know you do.

You stated; "Does this mean I should devalue this record as evidence of the marriage?".
If you re-read my comments, you'll see that I was simply making the point that a single record can support many different research questions and can't be analyzed out of context. It may be considered direct evidence for one research question, indirect for another and possibly negative evidence for a third. Does this "devalue" (your term) your record? Perhaps not when it's used to address whether there was a marriage. But, one can't say it has the same value in answering all the questions that one could raise about the marriage. That is why one can't really tag a single event or attribute with a level of evidentiary value. That value has to be assessed when one is looking at answering a specific question.

[AdrianBruce]

Adrian;
I believe you mentioned that you used all three tags; Census, Occupation and Residence. Could you elaborate on why you use both Census and Residence? ...

Yes. Somewhat as Helen says, I get Residences from several potential sources, such as parish registers, etc, so they generate Residence "events". Unlike Helen and some others, I'm happy to use the censuses as evidence of Residence - once I've discarded the visitors, etc, etc, that I've mentioned before. So I want to have all residence data accessible through one Fact Type - that of Residence.

So why do I still use the Census event? I guess that there are two reasons. Firstly there are those who are marked on the census as visitors, inmates, etc. They won't have a Residence event for that date - the Census event is therefore the only easily accessible record of where they were that night. (This may be an aspect that some are unfamiliar with - UK censuses from 1841 until sometime recently, document people where they were on the night of the census but the "relationship to head of household" includes terms like "visitor", so we can see those who are just passing through).

The second reason for still using the Census event is that it acts as a visible check whether I've actually got the census details for someone. Either I do a visual check that the Census event is present for someone in 1841, 1851, 1861, etc, or I can run a query that checks for the presence of the event in question when compared to dates of birth and death. (At least, theoretically I can run such a query - I can't remember if I do have one...)

Because the census events are just checks, I don't print them in narrative reports - although I'm not sure how well those with the relationship of visitors etc, show up in those narrative reports where census events are generally suppressed.

[Gary_G]

Thanks, Adrian. I now see what you do and why.

[RS3100]

That is why one can't really tag a single event or attribute with a level of evidentiary value. That value has to be assessed when one is looking at answering a specific question.

You can record the source (the census) and cite it against each event that it provides an evidential value towards. The assessment of its value and credibility is applied to the citation, so can be different for each event ("fact") that you cite it against. See the citation Assessment Dialogue in FH help.

[Gary_G]

You can record the source (the census) and cite it against each event that it provides an evidential value towards. The assessment of its value and credibility is applied to the citation, so can be different for each event ("fact") that you cite it against. See the citation Assessment Dialogue in FH help.

I am aware of the capability to assign different ratings to a citation, when it is attached to a "fact".

You will notice that I have consistently referred to only the Evidence metrics; Direct, Indirect and Negative. As the other quality metrics are an assessment of the source, they can be assigned to the source without knowing the research question. The Evidence metric is special. It is assessed on the basis of whether the source answers the research question on its own, requires other sources to do so, or possibly even acts as negative proof.

The problem with trying to assign an Evidence type to a source when attaching it to an "Event" or "Attribute" is that there is no clear non-trivial question being posed/implied. I suppose one could assume that the "Event" or "Attribute" is an assertion and your research question is a simple "Is it correct?". You could then assign an Evidence type based upon that. Is that what you were implying? If one does that, be aware that the Evidence type will likely need to change if you use it to support any other higher-level research question. Personally I don't see much value in assigning an Evidentiary type at that level in the process, because any research questions I've seen in standard research plans are almost never as simple as proving/disproving a single "Event" or "Attribute".

[Gary_G]

The thread seems to have wandered quite far from my original question "Is someone aware of an article or video that discusses the pros and cons of using a particular combination of the three?" I really thought it was a fairly direct question and hadn't expected a lengthy discussion. Thankfully; I did receive direct and relevant answers from Mike, Adrian and Helen. I think those will help. Thank you for that. Now I'm going to get back to my data entry...

[NickWalker]

The thread seems to have wandered quite far from my original question "Is someone aware of an article or video that discusses the pros and cons of using a particular combination of the three?" I really thought it was a fairly direct question and hadn't expected a lengthy discussion.

I agree that was a direct question in your first post and then you asked this question in the next one: "Could you clarify why you create a census record? I don't understand the value of being able to run queries on census entries, if the content can't be trusted." and that was what I and a few other people tried to answer and that's why it became a lengthy discussion.

[Little.auk]

For version FH 6 Calico Pie published a "Users Manual", written by Simon Orde, entitled 'Getting the most from Family Historian 6'. On page 61 it defines what is meant by "Fact", and I quote -

"In Family Historian, the word 'Fact' is often used as shorthand for 'event or attribute'. The Facts tab tab shows events and attributes"

As others have said - the term 'fact' is a very poorly chosen word - It implies "Truth", but one thing we learn very early in our family history research is that, even on original 'official' documents like marriage certificates, there can be quite a lot of fiction, both accidental and deliberate.

To be fair, this usage of Fact pre-dates FH - perhaps 'Evidence' would have been a better choice, as it does imply an element of unreliability and a need for evaluation.

A Census - this is one of the six 'main' events that occur in every person's life that provide the family historian with the base-line 'evidence' around which to build a person's family history.

If we applied the idea of not to recording a Census "because the content can't be trusted" to other Events, we would record very little!

I use AS which records census events for me this provides a quick answer to the questions "Was Fred Jones on the 1871 census?" or "Have I found Charlie Smith on every census he should appear on?"

I use the Property Box 'Census Tab' layout as per the attachment. This layout came from a previous post, that I am sure Mike will be able to point people to.

At the top it shows Birth year and Death Year, and listed below is the list of census events. So I can easily see whether I have records from all the censuses I would expect to find the person on.

It also shows an example of the most inaccurate fact on the census - the persons age!

Census Tab.jpg (128.07 KiB) Viewed 440 times

[tatewise]

I use the Property Box 'Census Tab' layout as per the attachment. This layout came from a previous post, that I am sure Mike will be able to point people to.

That Census Tab is from the FHUG Knowledge Base > Downloads > Property Box Tabs > Census UK but modified to remove the Occupation fields. Note that the Downloads image of the tab does not include the 1921 Census and the Age fields.
The update is derived from the How to add 1921 Census to list in 'Census UK' tab (20478) thread.

[AdrianBruce]

... As others have said - the term 'fact' is a very poorly chosen word - It implies "Truth", but one thing we learn very early in our family history research is that, even on original 'official' documents like marriage certificates, there can be quite a lot of fiction, both accidental and deliberate.

To be fair, this usage of Fact pre-dates FH - perhaps 'Evidence' would have been a better choice, as it does imply an element of unreliability and a need for evaluation. ...

Philosophy warning
Actually, as a mathematician who might be expected to be hot on terminology, I'm fairly relaxed about the use of the word "fact", on the basis that any "fact" in science is always only one set of experimental results away from being disproven - from becoming an ex-fact. Perhaps because of that being held as a fundamental principle, scientists tend not to worry about using the word "fact" - though by the logic expressed in the quote, nor are they enthusiastic about using it, I think.

Oddly enough, perhaps, I'm less keen on "Evidence" as an alternative term - "Evidence", to me, is what is in sources, and the way I work (and others don't!) the events and attributes are my conclusions.

You pays your money and you takes your choice....