details on census returns

skeptik4321 · Post by **skeptik4321** » 24 Mar 2008 16:18

Am I being too pedantic? If a census return has certain details about someone then surely that is what I should record? I have an ancestor who put a different place of birth on each census return, all within a few miles, yet the Gedcom Census programme took the birthplace I had entered in Family Historian. My inclination is to edit this manually, altho to do that takes away some of the advantage of the Autosource box. What do others think?

ID:2823

YouTube · Post by **Jane** » 24 Mar 2008 18:22

There is nothing to stop you using FH to enter multiple birth entries. Gedcom census only creates a birth record if one does not already exist.

ChrisBowyer · Post by **ChrisBowyer** » 24 Mar 2008 19:15

We tend to record the birth place from the earliest census if there's a choice on the grounds that the parents are more likely to remember... Often adults record where they were brought up rather than where they were born.

tommy166 · Post by **tommy166** » 24 Mar 2008 19:41

The auto-source is just a useful tool to help you record a transcription of the census. Most of the time you will need to modify some of its suggestions to match what is recorded on the enumerators form.

Whether you make separate birth events if the information recorded differs from that which you hold and then link them to the census source is up to you. However you would need to do that in Family Historian, as, like Jane says, Gedcom Census will only add birth details if none exist.
I tend to, like Chris, use the earliest recorded place, unless there is a general consensus in later censuses.

skeptik4321 · Post by **skeptik4321** » 24 Mar 2008 20:42

Many thanks for all the help, yes I shall take earliest recorded place or look for a consensus. It makes sense.
Val

JonAxtell · Post by **JonAxtell** » 25 Mar 2008 16:05

The point of the autotext feature is to take a lot of the grunt work of transcribing the census transcript. It can never produce an accurate copy of the transcript, so you nearly always have to do some manual editing. The fact that you have to doesn't detract from the advantage of having autosource fill in all the other stuff.

My census entries record the information off the census form, mistakes and varying names (including married names) and all. The birth event might sometimes never match up with what's on censuses.

As for multiple birth dates - when you have the problem of rounding dates on some censuses and people's lack of memory and other issues, the best you can do is put in an approximate date (with notes) for the birth until you have more accurate information.

jmurphy · Post by **jmurphy** » 25 Mar 2008 16:12

Yes, you should record with each source (not just the census) the data which is on the source. The auto-text box filling in some data is just to save the user a bit of typing -- Nick makes it clear in the help that the auto-text will have to be tweaked to fit what the census actually shows.

You can use the up/down arrows to put your 'preferred' birthplace at the top of the list in Family Historian, but you should still enter all the variants. The whole point of having the computer is so it can keep track of all those variants for you.

The census in the US, especially the modern day ones, can have many more questions than what is shown on a UK census return, so I generally correct the GC auto-text to reflect the data that would show on the UK census and then enter anything else in Family Historian, using the Auto-Source Citation feature.

I should also transcribe each census fully for my records, but so far I have been too lazy to do so. Tools exist to help with this (for example, there is a shareware set of Excel spreadsheets called 'Census Tools') but I haven't knuckled down to do it yet. [oops]

My excuse so far is that because I haven't done transcriptions, it forces me to go back to the images, and when I do, I find new things each time.

Jan

ChrisBowyer · Post by **ChrisBowyer** » 26 Mar 2008 04:09

jmurphy said:
you should record...

Never forget you're doing this for fun... you don't have to record anything, you can do whatever suits you.

But more seriously, I try to avoid transcribing anything I don't have to. Every transcription introduces errors. If you ever need to refer back to a census (for example) to see what it really said, you'll want the original not your transcription. And isn't it a shame that all we have is the enumerators transcription with its (sometimes obvious, sometimes not) mistakes... If only they'd kept the householder's forms.

ChrisBowyer · Post by **ChrisBowyer** » 26 Mar 2008 04:25

JonAxtell said:
As for multiple birth dates ... and people's lack of memory and other issues...

It is surprisingly common (try it on your friends if you don't believe me), if you ask a woman how old her children are she'll say something like 'he'll be 14 this year', i.e. one year older.

Post by **NickWalker** » 26 Mar 2008 11:59

ChrisBowyer said:
But more seriously, I try to avoid transcribing anything I don't have to. Every transcription introduces errors. If you ever need to refer back to a census (for example) to see what it really said, you'll want the original not your transcription. And isn't it a shame that all we have is the enumerators transcription with its (sometimes obvious, sometimes not) mistakes... If only they'd kept the householder's forms.

I believe for the 1911 census we get the original householder's forms.

When I first developed and started using Gedcom Census the census images were not widely available, Ancestry, etc. hardly had anything on line and I had hundreds of census entries that I'd previously recorded as notes in Family Tree Maker which I then copied and pasted into the Gedcom Census source text. I've continued to transcribe the census entries as much for consistency as for anything else and I do agree that it can add more potential transcription errors. Having said that this happens at all stages in family history recording, whenever you record any kind of event or attribute you might be typing in the wrong information.

Thinking about it I could misread information on the microfilm reader at the record office (potential error 1), write the information down wrongly on my piece of paper (potential error 2), misread it when I get home (potential error 3), mistype it in (potential error 4). This ignores all the previous potential mistakes the vicar or priest might have made when recording the information in the first place! Its a wonder we get any accurate information at all

I find it useful in Family Historian to be able to quickly see at a glance who was in a census household by looking at the source text, it takes much less time than viewing the census image, finding the family on the page, maybe looking at a couple of pages, trying to read the hand-writing, etc.

I completely agree though that you should record things the way you want to. I've tried to make Gedcom Census very flexible by offering various options, including 'recording method 2'. Having said that, beginners often want advice as to how to do things and we should always be happy to offer advice.

jmurphy · Post by **jmurphy** » 26 Mar 2008 16:47

ChrisBowyer said:

jmurphy said:
you should record...

Never forget you're doing this for fun... you don't have to record anything, you can do whatever suits you.

But more seriously, I try to avoid transcribing anything I don't have to. Every transcription introduces errors. If you ever need to refer back to a census (for example) to see what it really said, you'll want the original not your transcription. And isn't it a shame that all we have is the enumerators transcription with its (sometimes obvious, sometimes not) mistakes... If only they'd kept the householder's forms.

Chris, your point is well-taken. I did not mean to say 'do it this way or you are doing it wrong' but rather, 'it is better practice to write down what is actually on the record'.

For instance, when I found one census record, the surname was mis-indexed by Ancestry. I downloaded the original image, but what happens later if I discover I need the pages on either side? If I forget that the index is messed up and I try to search for that family again with the proper spelling, I won't find them, so I'll have to search for one of the other families on the page, and so on. But if I put in my notes that the index is messed up and make a note of Ancestry's hideous index record, the computer can remember why I couldn't find the record with an ordinary search.

I totally agree that we should have fun, but it's no fun to shoot yourself in the foot, and my remarks were intended in that spirit.

Jan

jmurphy · Post by **jmurphy** » 26 Mar 2008 17:00

A quick comment on what Nick said:

There are some UK census years for which I do not have images yet, and I only have data from FreeBMD.

I'm embarrassed to say that it took me quite some time before I figured out I could copy the information from FreeBMD's transcription and paste it directly into the Gedcom Census source text box. I was using Nick's Auto-text feature and hand-correcting what GC does, rather than thinking about what I was doing. [oops]

I mark these as 'census - transcription' vs. 'census -- image' (or 'census -- index' for the ones in which I only have someone's index record) in the source type field to remind myself which ones are which.

Jan

ChrisBowyer · Post by **ChrisBowyer** » 26 Mar 2008 17:33

I think we're all right but with a slightly different perspective... I wasn't trying to be argumentative, and I do find I have to remind myself sometimes (when inwardly wingeing about the amount of time and effort that goes into the tree) that we choose to do it.

But I think whether you embed or link to (which is what this boils down to) any information is a broader discussion than just census entry. Indeed, whether it is a source transcription or anything else, which you do is a matter of balance depending on the individual circumstances, most notably how easy it is to refer to the original, rather than a matter of right or wrong.

Generally we put transcriptions in 'Text from source' only if it seems to say something different from the data, otherwise simply to record that we got this information from a parish register, or a marriage index, or a censes, or whatever, with references that can be checked if need be, seems sufficient.

Incidentally Ancestry (unlike FreeBMD who make you feel like Oliver Twist if you try to sugest a possible correction) encourage and thank you for it, and I've had many an occasion where I've thanked myself for correcting one ages ago that I would have had a hard time finding again (and it makes you feel good).

jmurphy · Post by **jmurphy** » 26 Mar 2008 19:16

Chris -- I thought we were discussing, not arguing. No hard feelings here!

Yes, we do have different perspectives, and this is one of the things I enjoy most about having this forum for discussion.

IMHO, it is useful to have everyone say 'I do it this way because...'

Whether the person who asked the question chooses my way or your way or someone else's is not important. What is important is that the questioner has a more informed basis to make the decision. The solutions not chosen are just as important to that process as the solutions which are.

You make a good point. If one asks the question 'what do I want to enter in Family Historian' one answer might be 'everything I want to be able to refer to in a hurry'.

The trouble is, when one is just starting out, it is difficult to know what that information is.

Jan

Post by **NickWalker** » 26 Mar 2008 19:28

Getting back to original point that was made. Does anyone think Gedcom Census should have the option to record additional birth events? I've always steered clear of this as it is likely to cause confusion for others who see my data if individuals have more than 1 birth recorded but it could be done. I also assumed it would be more likely to cause problems when transfering data to Genes, Ancestry or other programs.

I could add this option but it would take quite a bit of work to do it properly and would only be worth it if it was likely to be widely used. Do any of you use multiple birth events for individuals? Opinions welcome.

JonAxtell · Post by **JonAxtell** » 26 Mar 2008 22:39

My view would be no, GC shouldn't create multiple birth events. If GC was instead a tool for the entry of BMD information, then there could be an argument for it, as such sources tend to be more accurate. But as census data is not accurate, potentially creating a birth event for every census would be unweidly and not provide any benefit.

The main thing to do if you come across conflicting ages on censuses is to record your thoughts about why you know that they relate to the same person (eg. family group), note any possible census rounding issue, speculate that they lied about their age, etc. Creating birth events makes it hard to keep track of this as programs don't allow you to easily see the conflicting information. The best solution is to record all such information in a note.

The general principle I follow is to create structured data where there will be a benefit to do so. This could be for analysis, querying, reporting, etc. There comes a point where structuring your data to the nth degree gives diminishing results. For instance, I don't beleive there is real need to record every single educational achievement using the EDUC attribute. Unless you've got a lot of people in academia there won't be much analysis or querying possible, and the narratives produced are always a bit dry from such basic data. Such information is better documented as prose in a note attached to the individual. Similarly, there is no real point in using the extra fields in the name structure to record first name, surname, title, prefixes, and suffixes.

tommy166 · Post by **tommy166** » 27 Mar 2008 01:09

I only ever enter a single birth event for any individual. I base it, where possible, on census information and try to find confirmation in the GRO indices and/or parish baptismal records. All of the census events have source text and linked images so I can see how I have drawn my conclusions. If necessary I will add a note to explain my reasoning.

I would find having multiple birth events more confusing than useful. Having Gedcom Census create or complete birth events is useful; but if there were always an option to add the birth event I think it would make the procedure unnecessarily complicated. It is, after all, possible to add another birth event in Family Historian and link it to that census source if you wish to record all the variations separately.

jmurphy · Post by **jmurphy** » 27 Mar 2008 06:20

I think it is important to remember that we don't always know who reported the data to the census-taker. People often say 'so and so lied to the census taker' but we can't know who told them that age. For all we know, someone told the census taker 'he was born in 19xx' and the age was calculated from the birth year' -- and as already discussed here, it is common for people to give the age someone is about to be in response to that question.

So for me, a birth year from a census record is iffy to start with -- and should only be the primary information about a person's birth where no other record exists.

I do use multiple birth events. I don't upload my data to any online service.

As I've mentioned elsewhere, I use FH as a tool for analysis. If one source says a person was born in 1900 and another source says they were born in 1902, I have two events so I can look at them in FH and have a record of which sources says which date. I also have custom events for birth registration, marriage registration, and death registrations, so I can keep track of that as a separate date.

It's very easy to go into Family Historian and add information -- I do this for US census data in any case since there is so much other data that needs to be recorded. So while I do appreciate the feature that will add in a birth event if one doesn't exist already, I don't see the need to have Gedcom Census add the extra birth event for me if one already exists.

Jan

ChrisBowyer · Post by **ChrisBowyer** » 27 Mar 2008 06:41

Multiple birth records seems a strange idea to me... A person is born once; what you have is multiple, possibly conflicting, references to that event, which surely belong in notes, sources, text from source, etc., with the birth event itself being your aggregate opinion as to the most likely answer.

ADC65 · Post by **ADC65** » 27 Mar 2008 09:49

It's not a function of GC I'd use, but it's good of you to offer it Nick. Like others here I prefer to only have one birth event; I then record any additional information or discrepancies in an attached note.

jmurphy · Post by **jmurphy** » 27 Mar 2008 16:11

ChrisBowyer said:
Multiple birth records seems a strange idea to me... A person is born once; what you have is multiple, possibly conflicting, references to that event, which surely belong in notes, sources, text from source, etc., with the birth event itself being your aggregate opinion as to the most likely answer.

I see your point, but here is my reasoning.

I was taught to mark the source my information came from, and even the session, so that if you find out later that a source is unreliable or for some reason the information from that session was unreliable, you could go in to your data set and pull all that data out and toss it.

Since I am using the computer as the modern-day equivalent of my card file, I am trying to mimic the separate cards that I would have if I were doing this on paper. Each source would have its own card. This makes it easier to disentangle things if I find out I have sorted them out by mistake (e.g. if I have merged two records and find out later they are two separate people).

I can certainly see the value in doing it your way, and if I were using FH to produce reports and to publish a web site, I would have a cleaned-up database which had only one event, to do just that -- although I think I would keep my custom BMD registration events in any case, because to me the birth and its registration ARE two separate events, just as a birth and a christening, while close in time, are not the same event.

Jan

PatrickT · Post by **PatrickT** » 28 Mar 2008 15:20

I always try to stick with one birth event based on the most reliable source. If other sources differ I still cite them but add a local note to the birth event stating how they differ.

Tombaston · Post by **Tombaston** » 30 Mar 2008 10:09

I also use multiple birth events. If you have someone on three censuses they can easily have three different calculated birth years and sometimes different places as well (I regularly find this with locations in the east end of London). If I don't have their birth certificate I have no way of knowing which is the most accurate, so I need to record them all.

As with many things in family history there is no wrong or right way to record things. Decide which method you prefer and stick with it.

Tombaston · Post by **Tombaston** » 30 Mar 2008 10:17

Nick Walker said:
Getting back to original point that was made. Does anyone think Gedcom Census should have the option to record additional birth events? I've always steered clear of this as it is likely to cause confusion for others who see my data if individuals have more than 1 birth recorded but it could be done. I also assumed it would be more likely to cause problems when transfering data to Genes, Ancestry or other programs.

Sorry Nick, I missed your question. Although I record as different events, if I find a new census which suggests a different birth year (or place) I add it with a quick edit in FH. So I use GC to add the new census, making changes as needed in the Auto Text and save with the image. I then go to the individual in FH, copy the source from the newly added census, create a new birth event and paste the source to it. This can be done fairly quickly, so you don't need to add the facility to GC for me.

andrewbraid · Post by **andrewbraid** » 30 Mar 2008 16:56

I have often thought that it would be useful if FH had an additional qualifier for a date of birth, namely 'Derived from Census data' or something similar.

I never know, when census data is the only source of a birth event, which of the three existing qualifiers to use. For consistency I always use 'Calculated' but I also use this when a birth date is worked out from an age at death and the latter is generally, but not always, more reliable (at least there is only a single event to calculate it from).

However, I suppose that such an idea would not conform with the GEDCOM Standard. Pity.

Family Historian User Group