Should we be creating Census categories?

Question

Until I read this question, I didn't realize that we had Census categories. So, I checked and discovered that, for Chatham County, North Carolina, there is only one category (1830). Is the intent to have categories for all Census decades? Or is this something special?

A brief review looks like there is no consistency at all the naming convention for Census categories, so I would be hesitant to create one until I know what the standard is.

Comments

There used to be a moratorium on the creation of US census categories, dating from 17 July 2021:

https://www.wikitree.com/g2g/1271442/please-discontinue-adding-new-us-census-categories-for-now

On 19 March 2022 it was questioned whether the moratorium still held:

https://www.wikitree.com/g2g/1390945/census-categories-still-on-hold

By 5 July 2023 it appears to have fallen into abeyance:

https://www.wikitree.com/g2g/1603598/are-we-creating-any-new-census-categories

It seems desirable that when there is a moratorium on creation of categories of some kind, it be of limited duration. If after a reasonable delay a new category structure has not been proposed, people should be free to go ahead and create categories in accordance with their own best judgment.

---

It seems desirable that when there is a moratorium on creation of categories of some kind, it be of limited duration. If after a reasonable delay a new category structure has not been proposed, people should be free to go ahead and create categories in accordance with their own best judgment.

There has never been a proposal and approved structure for these categories. From the Help pages:

"Do not create a category unless it fits neatly within an existing hierarchy and uses an agreed-upon category naming scheme."

While I understand your viewpoint in the moratorium, users should not really "go ahead and create categories in accordance with their own best judgment." They should instead use the process outlined above by starting a discussion in G2G (which is happening here).

---

But there is already a huge existing US census category hierarchy in place, containing hundreds of categories with some sort of naming scheme. See

Category: United States Census Records

and its subcategories. Why can't a missing county category for example be created, in a way compatible with what's there already?

If the Categorization Project had indeed made a proposal within the next month or so after 17 July 2021, and agreement has been reached, everything woud be straightforward. Apparently that was too daunting for the project, and it would be all the more so for individual users to attempt now. In its absence, continuing with the existing structure does not make things qualitatively worse; it merely increments an eventual cleanup if one is eventually decided on. It seems a pity to leave the categorization of probably thousands of new profiles hanging in the void indefinitely.

---

Jim, the Categorization Project was not working on the proposal. This was being undertaken by another member. And I did not mean to suggest that they cannot use or create new categories - just that a blanket statement like "people should be free to go ahead and create categories in accordance with their own best judgment" is not the correct way to handle things.

---

While I concede that there shouldn't be anarchy, I stand by my point that there should not be indefinite moratoria on category creation. This is not the only instance of that.

Answer 1

Categories should not be created unless there is a profile to go in them. So for your Chatham County, North Carolina example it only has one years of census categories. Obviously the other years either doesn't have any profiles identified and someone wanting the category.  A different example might be https://www.wikitree.com/wiki/Category:Carteret_County%2C_North_Carolina%2C_Census_Records which has what looks like every census year (That is sourceable) category created with profiles attached.

The format for the categories that profiles go in is quite simple for the Federal Census categories using Chatham County as the example

1830 US Census, Chatham County, North Carolina

• The year of census   eg 1830 US Census
• The location              eg   Chatham County, North Carolina

So for a 1840 census in Chatham County, North Carolina would be [[Category:1840 US Census, Chatham County, North Carolina]]

Comments

I have enough Chatham County profiles to create categories for every census from 1790 through 1950. But ISTM there should be consistency in the category title structure before I proceed.

It looks like the category structure is: United States Census Records => (State) Census Records => (County) Census Records => (Year) US Census Records, County, State.

But then there appears to be a parallel category structure: Year United States Federal Census => Year United States Federal Census State  => Year US Census, County, State

Are we supposed to use both of these?

When I click on edit, the Chatham Count 1830 Census has this:

[[Category:Chatham County, North Carolina, Census Records]]

[[Category:1830 United States Federal Census, North Carolina]]

So that first category seems to introduce yet a third element into the categorization.

I'm really confused.

---

The [[Category:Chatham County, North Carolina, Census Records]] connects to a County category and the State overall census category. That is because if people look at the county they could find census categories related to the county from there. And it is the same from a State census category. 

The [[Category:1830 United States Federal Census, North Carolina]] connects to the United States census category for that year and the state census category in a different way. 

For profiles all that matters is the destination category. Then the parent categories are quite easy and are more likely to already be created. Those categories also have a simple format if you are used to them. It is also a consistent format. It is also easy to copy from another category either from the same county or a different county in a state and just change the year or county information to suit whatever year and county you require a category for. 

Answer 2

In the UK, we discourage creation of census categories. A census (should) contain the entire population, so each category would contain (theoretically) millions of people.

Comments

“England” also contains millions of people, yet it is a high-level category. A census of an English village would be small enough to have a category of accepted size. I don’t see any reason to discourage a census category structure that complete mirrors the place name structure.

---

'High-Level' categories are not supposed to contain person-profiles.

---

This is just an example:

Category: 1841 census (there were over 18 million people)
Category: Devon, 1841 census:(over half a million people)
Category: South Pool, Devon, 1841 census (555 people, one of the smaller villages); or
Category:Exeter, Devon, 1841 census (31,312 people)

Can you see how unwieldy and pretty impossible this would be? And the population mostly grew in the later censuses!

---

The problem with the analogy Ros is that the equivalent United States Census category for Devon (1841) would be something like New Hampshire (1840). An English County is not the same as a United States County but is an effective level above (A US State). The profiles don't go in a State level census category and it wouldn't work if someone expanded the use of census categories for England or the United Kingdom to have profiles in a English County equivalent category. 

Not that I support this but there are some people who have created or used census categories related to England or United Kingdom Censuses. Some of the categories were made this year even. So some people must find value in a census category. 

The argument for or against any category should never be there will be too many profiles in it but should be how will this category be useful for people. I could make the argument that cemetery categories are useless as you could just look at the sources to see where they were buried or a Location category is useless as I could just search via Wikitree + and find the profiles with Birth/Death or Marriage locations in that location. Migration, Occupation, Military, Art, Health, Unusual Experiences could all have arguments why they are not needed or that they are needed. The United States Census categories are mostly already existing with the parent category structure in place and just some more specific time/place categories not made because the profiles that could go in them are either not identified or the PM does not like categories which is their right. Some people do find the United States Census Categories of use even if others do not. 

---

If census categories mirrored place name categories, there would be no profiles in the Devon, 1841 Census category, just as there should be none in the Devon category for the place itself. Both categories should contain subcategories, and subcategories of those, etc.

Any locality in England small enough to have profiles in the corresponding category would be such that the corresponding 1841 census could have profiles in it. In fact, the latter is naturally a subcategory of the place name category and would always be smaller. Like, Littleham (Exmouth), Devon, 1841 Census should be acceptable size if the larger category Littleham (Exmouth), Devon already is.

Now if we don’t want a hierarchy of census categories and would rather just have census categories as subcategories of the corresponding bottom level place names, that is fine with me. But that’s not the same as discouraging census categories altogether.

---

In my example, I was only trying to show the huge numbers.  If you know anything about creating categories, you would know that the imaginary Category: 1841 census would not contain any profiles, let alone 18 million, and the imaginary Category: Devon, 1841 census would not contain any profiles, let alone half a million.  But look at Exeter - over 31 thousand profiles!

Answer 3

Why?  The location categories for your profiles already would group them.  Don't see the advantage of doing census categories at all, since the censuses are available online for the most part prior to certain statutory dates.  Adding censuses by year as category is redundant.  We're not the census bureau.

Comments

Because people move. A location category is longitudinal, while a census is cross-sectional. A census category will pin down when a person resided in a certain place, while the location category does not. And it is genealogically useful to see who resided in the same place at the same time.

---

They can be a bio item without being a category.  I use censuses sometimes for recent profiles, in the bio.  They are an aid in figuring out who is still around or what children there may be for whom BMD data is not freely available due to privacy restrictions or lack of indexing.

The only censuses I use consistently are the very early ones for New France, of which there are in fact very few nominative ones.  And the population base is quite small, the biggest census being 1681 census, which runs into 30-32k total polutaiton, broken down by areas.

---

The problem is that some people feel that they should only use a Birth Location as the only location category or the Birth/Death/Marriage as the only categories. So it could be quite easy to miss that a person was not in those locations when a census happened.

The Census categories for the United States help to connect people who where in a certain county in a certain year. A Location category groups people that lived in a location whenever the location was around. It doesn't group them in a purely genealogical sense as it could contain people hundreds of years apart for some locations. A census category could make it easier to figure out certain families lived close by and when whereas a location category only tells you they lived nearby. So the Census categories help to narrow down people in a location.

Any category could be waved away as a "Bio" item including Location, Census, Military etc. Does that mean that we don't use them, of course not. Everyone uses categories in a different way to everyone else. I for example do not use any Migration categories yet a lot of people do use them. I know several people who remove cemetery categories from profiles they manage as they do not want them on the profiles. Are those reasons to prevent further Migration categories or Cemetery categories being made. No, it is not. As long as the structure is there that new categories could be slotted into the hierarchy then those categories could be made. (With profiles attached).

---

The problem with this logic is that you need to have all or a majority of the profiles of people named in the census included in the category, or nearly so, for it to be truly useful.  Which is a whole lot of work identifying / creating .. profiles.  Repeat ad nauseam for each census.

---

The problem with this logic is that you need to have all or a majority of the profiles of people named in the location included in the category, or nearly so, for it to be truly useful.  Which is a whole lot of work identifying / creating .. profiles.  Repeat ad nauseam for each location.

This statement is as correct as the statement you made about census categories. It does not mean that we should not add a profile to a new location category unless we created every profile possible that also interacted with that location. It is also impossible to know if we would every have every profile for a location as sometimes there will be no sources proving it. 

For a United States County census category there is a finite number that would be in that category however small or large that number can be. There is also a clear division by year thus meaning the categories would not get overwhelmed with every profile that was in any census being in the same category which they would end up being if just by location. We also have a source proving a time and location which would support a category. I see so many profiles with categories on them with nothing to support the actual category. 

Categorization works best when we encourage people to categorize in ways that work for them. Discouraging a certain type of category because some people do not like them is not helpful in encouraging people to use categories overall. After all a Location category on a profile might lead to a census category which could lead to a migration or occupation category. And vice versa. 

---

---

Danielle: that’s not a problem of logic, but of time, of which there’s plenty. Here’s a transcription of the 1774 census of Bristol, Rhode Island with almost all heads of household linked to profiles:

https://www.wikitree.com/wiki/Space:1774_Rhode_Island_Census_Project:_Bristol

Here’s the corresponding category:

https://www.wikitree.com/wiki/Category:1774_Census_of_Rhode_Island%2C_Bristol

---

As is often pointed out, categories are there to group people.  If somebody adds a category to a profile without anything in the bio supporting said category, then there is something wrong.  Categories are NOT meant to replace bios.

Answer 4

One of the discussions that we usually have, when talking about categories, beyond the how of creation and organization of categories, is the why of creation of categories.

There are many here who focus on use of categories, and stickers too, without a strong argument on why certain categories should be created. The usual argument is that categories should be used for some genealogical purpose. A good example of this are cemetery categories. By categorizing profiles into specific categories, they can be searched to look for people who are related to other people.

But some categories seem to be used, not so much for genealogical purposes, but much like some stickers, as just a way of noting things that should be in the Biography. For example, some of the categories regarding an occupation seem to serve less a genealogical purpose then just biographical notation. See, for example, the category of Farmers. I can't really use this for genealogical purposes, like finding other related people. I can find other Farmers, for sure. But I would be better off using the FAN Club or Cluster genealogy to further my research.

So where does this leave Census categories?

While it could be argued that a Census category structure could be used to search for related people (as related people can sometimes be found near each other in a location). However, the only way to make this useful would be to create all levels of subcategories of Census records (each decade, each state, each county, each enumeration district), and to systematically add every person, to every relevant category. It would be a huge undertaking. 

As Ros said, it would potentially contain millions of people. As Danielle said, location categories already group people. And we're not the Census Bureau. This is the reason for the hesitation to continue with creating and using these categories. It seems sufficient to put Census citations on profiles, use location categories if you want. It's my opinion only, but categories should be used sparingly.

Comments

After all of this discussion, I'm inclined to agree with you. So, I won't use those categories nor will I create any new ones.

---

While I find some WikiTree categories quite useful in genealogical research, I can't think of a single instance where I used a WikiTree census category as a research tool.

Answer 5

"The average number of counties in a US state is 62." The range is 3 to 254, and we have 50 states... Factor in the census years, and we'd end up with quite a few categories.

Migration, place name studies, predictive associations of surname by place, and identifying completeness might be valid recourse to justify adding people to the county level by census year categories, but is anyone really doing that? It would be a tremendous undertaking.

On occasion, I have researched which census years are missing and found the surname incorrectly indexed, so I can see where ticking off the located census sources by a category might be useful in that instance. I don't recall ever actually using a census year/county category methodology in that process, however.

On the other hand, locating and picking a category location name match has gotten harder with the growth in the decennial category names alphabetically preceding the simpler location name. I make an effort to include all event location categories to a profile and personally think that's sufficient.

Comments

After following this discussion, it is my opinion that WT ought to consider no longer using census categories. They don't seem to serve any useful purpose, and the size of the category would be immense.

Of course that's not my call, but, as I said earlier, I will not be using them for the reasons so well-articulated here.

---

It's apparent from the discussion that when the subcategorization is fine enough the sizes need not be gigantic, and that some people do find these categories useful. They shouldn't be compulsory, but they shouldn't be forbidden either.