Looking around for my Miller ancesters in Licking county I've made some analysis on why I don't quite like things as they are at present. I mean, I do like that it's not too expensive to subscribe to Ancestry or someone else for access to census images and some indexes, but they only take you so far and I constantly get the feeling I'm re-inventing the wheel every time I start tracing things down.

Fact is there are tons of people researching ancesters even for one county and if they got together and did things once and for all they could make it better for all in the future. Are there gigantic problems to creating family trees based just on the census records? Let me give a quick overview of what I have in mind.

I'll do Monroe Twp since that's one township where I have lots of ancesters. Typically the census records have 20-50 pages. Each page might have 15 families averaging 3-4 individuals. If we're talking all the census from 1850 to 1930, i.e. all the ones which contain everyone's name, we'd have about 400 pages or so. If we had 10 people each commit to taking a page a week and adding it to a tree (using census years and pages as "ancesters" of the individual families, the entire corpus of Monroe Twp census data could be available as a GEDCOM in less than a year. The mechanics of say when you could link a family from one census to that in another census would have to be hashed out, but even a fairly badly done effort might save lots of people lots of time. What does everyone think?

You must not be familiar with Genweb sites. They are trying hard to have info out there free. Some of what you want may already be there. Page down for census info. Contact the list administrator for that site if its not there and you want to contribute.
http://www.usgenweb.org/projects/projects.html
I have just completed only 1/2 of 1890 Licking and its in the process of being proofed.I also did Summit Co 1890, not sure of that status yet.
Some info has been put on under various states and their counties. Look around first before you duplicate. Their is also a census site online, can't think of the address just do a search for US Census and see what turns up.
Barbara

You're right I wasn't aware of their projects on the census. But that was partly why I posted the message, hoping someone would tell me of something like this. But I'm surprised I haven't been linked to it before. Anyway, I need to go back and see what exactly it consists and whether it does what I'd like. The big thing I'd want would be being able to find the same family in consecutive censuses without having to do a lot of work.

Any, thanks much for the tip.

http://www.usgenweb.org/statelinks-table.html
The above is the main site, I urge people to use it. Info is free and they do ask for vol. to submit, not all sites have what you what but often there are other links and address'. Some list the Historical Soc sites that have info also.Enjoy it.
Barbara

You stated:

"Are there gigantic problems to creating family trees based just on the census records?"

I can think of one major problem with this, at least for the census records before 1880. They were not proof of relationship, only of residence. If you were to assume every child in a household were siblings you would be making a grave error. Cousins, nieces, nephews and grandchildren often lived in a household and people assuming they are the children of the head of household are the reason we have so many bad Gedcoms floating around.

The later census records improved greatly by adding the relationship to the head but until they started adding the age at marriage and the marriage sequence it was also impossible to tell whether the wife was the mother of the children listed. Even then without something explicitly saying the relationship to both, it is not proof.

Census records are not a primary source of information. They are wonderful starting places but use them only as a secondary source to back up other sources. Sometimes secondary sources are all you have and a preponderance of them can sometimes be used in place of one good primary source but only if they all agree and not all census records do.

The other problem is that many census records can be transcribed many ways. Two people transcribing the same page may come up with many differences. Transcribed census records again are a good starting point, but you should always try to view the image yourself and make up your own mind.

I applaud you for your ambition but there are already census transcription projects being attempted. They are not creating large databases as you suggest but I think that is something that should be left to the individual researcher. I have seen only one attempt by an individual to create a database of all families in an area and that was the Tavenner Collection which is on file at the Balch Library, in Loudoun Co., VA. Mr. Tavenner used tax, marriage and baptism records to track the population of that county from the early 1700's to his present and spent a lifetime in that effort. It would take a considerable time and effort even with today's primary resources to do the same. Secondary sources may be quicker but no where near as reliable for the reasons I mentioned. Few of us have Mr. Tavenner's relentless enthusiasm. I wonder if he would have gotten as much done if he lived when there was TV?

I agree with what you say, but I'm not asking for creation of a fault free database, but one which can be used to trace ones way through the existing maze of data. If you have a Mary Green age 1 in the 1880 Licking Co. Census wouldn't it be nice to be able to turn to the same Mary Green in the 1890 and 1900 censuses without having to search through 30-60 pages of census sheets, even assuming the family stayed put in the same township? And if someone knows who this Mary married wouldn't it be nice to be told just where to look in 1910, 1920 and 1930? Even if there's some doubt between 2 or three potential husbands, listing them in the notes for this Mary Green would allow one to quickly find them and see which one fits information you've previously had.

And my point is precisely that one doesn't have to do an entire area by ones self. A relatively few individuals can take on a limited area and create a set of page transcriptions. A person could take charge of the output and combine it into one set for each census and then the fun begins as people report the links they know.

The initial links would be Census year => State => County => Township => [ ED ] => page => family. If a family is found to be the same in two consecutative censuses the link up or down is made initially by noting the location of the family in the other census. After a large number of such links are made, including information from other sources which are noted (Marriage to Fred Ashbrook 18 Apr 1906 - Newark Court house records + location of Fred Ashbrook in that or another Census), then a compler combines all of them into a master database leaving the initial node for the first census a family head appears in and noting the census location for other years and for descendents. A major undertaking, yes, but it would only have to be done once for all time. Thereafter volunteers could be in charge of verifying changes and additions Now, as I said, this can be done starting at the lowest level, the township or enumeration district and then larger groupings can be combined as desired. Assuming that any known linkages outside the township are already in the individual township records, it's a simple matter for a county coordinator to scan through and link up the data.

OK, I'm probably oversimplifying, but I'm still just at the overview / think-it-out stage.

Some census records for Licking Co are online
http://www.rootsweb.com/~usgenweb/oh/licking/census/
and http://www.rootsweb.com/~ohlickin/#links
and hopefully my 1890 Veterans Census will be up soon.

Barbara

As I pointed out to someone else off-line, there's a difference between the Census images and a transcript of them. I have access to all the images, it's converting that data to something your computer can understand that's the trick. The purpose of having the transcripts, of course, is to make it possible to refer easily back to the images, and to use the information to build family trees.