Terry's TMG Tips

Terry's Census Tag
for 1790 to 1840 US Census

This page created 26 Feb 2007

Applies to Version 6 & 7

This article describes the Census Tag that I have developed for the 1790 to 1840 U.S. census - the years in which only the head of the household was named. I expect the techniques could be adapted to any other census that provides counts of household members by group rather than listing each member by name.

My article on Recording Census Information describes my approach to recording census information, which is aligned with the Census tag I describe here, and some important caveats about its application. Two companion articles describe my Census tag for 1850 and later, and a simplified version some users may prefer.

The U.S. Census for 1790 through 1840 produce special issues because only the head of the household is named, with others simply counted by sex, age bracket, and free or slave. Thus one cannot be sure who was actually in the household, other than the named head. But in many cases one can make an informed assumption, and my method allows recording those conclusions.

Some users get all fussy about including detailed warnings for their readers that we really don't know who the other members are, and are just guessing. I use a simpler format that I think adequately conveys the uncertainty, but users will have to decide for themselves how clear they feel they need to be on this point.

What It Does

The Census Tag described below records and provides narrative output with the following information:

For the head of household:

• Census title, e.g. "1840 Federal Census"

• The municipality, county and state of residence, to the extent provided by the record

• Lists the number of persons in the household by sex and age bracket, and identifies them as spouse or children of the head of household, to the extent information from other source can provide a basis for that identification

For others believed to have been counted as members of the household:

• Census title, e.g. "1840 Federal Census"

• Municipality, county and state of residence

• Name of head of household

• Relationship to head of household

This provides as complete picture as possible of the household in the narratives about the head of household, as well as key information in the narratives about other members. This works well in narrative output to Second Site where each person's narrative stands on it's own, but at the price of some repetition in Journal narratives when the whole family may be listed together.

Dealing with Relationships Between Household Members

With members of the household not listed by name, one can only compare the numbers of males and females by age brackets with the known or estimated ages of the spouse, ages and marital status of children, and any others who might have been present in the household. I find in many cases at least some of the household members can be identified with reasonable assurance.

There is a virtually endless list of possible relationships of members of the household have to the head of household – daughters-in-law, sons-in-law, with or without their children, mothers, fathers, brothers and sisters, grandchildren with or without their parents, uncles, aunts, and many more. But as a practical matter, I find that I can rarely make a sound judgment about about any but the spouse and children, and occasionally, an aging parent. When there is no reasonable basis for making an identification I declare them "unknown" or on rare occasions, include a guess as "possibly..."

Dealing with Groups Household Members by Age and Sex

The Tag deals automatically with:

• Having one or more than one person in each age/sex group.

• Identifying each unnamed person as the spouse, son or daughter.

This is done with the use of Roles. Probable household members are assigned Roles which reflect a combination of the believed relationship to the head of household and the age/sex bracket, for example:

• wife 25-35

• son under 5

• daughter 5-10

Of course the actual age brackets differ by census years, as the various censuses used different brackets. Note that one could be more politically correct and call the "wife" Role "spouse." But I've never seen a female head of household in this period where I believed her husband was present in the household. And doing so requires creating both male and female sentences for this Role, which seems to me like a waste of effort, so I don't recommend it.

In order to produce the correct wording for the various members of the household the system must account for any given sex/age bracket containing only one or more than one person. That is done by having "pseudo-plural" roles like these:

• wifes 25-35

• sons under 5

• daughters 5-1

Note that this is not necessarily a real plural - that is, "wifes" does not mean we think there were really two wives. Rather, it means there are 2 or more people in that age/sex bracket, one of which is being identified as the wife. (I like the "incorrect" spelling "wifes" as a reminder that this doesn't really mean there were two wives.) Similarly, the role for "sons" may be used when only one son has been identified, with the others in that bracket not identified ("unknown"). Choosing the plural role rather than the non-plural one causes wording like "...was probably one of the two females age 25 to 30..." instead of the wording "Was probably the female age 25 to 30..." in the output. The details below may make this clearer.

Sentence Structures for 1800

Below are the Roles and corresponding Sentence Structures I use for the 1800 Census Tag. Those for 1790 and 1810 through 1840 are similar, but have somewhat different roles to accommodate the different age brackets used in those years, as listed at the end of this article. This system requires custom Sentences for the standard Principal and Witness Roles, and the creation of the custom Roles listed below.

Role used for the head of household:

Roles use for wife of the head of household:

Roles used for children of the head of household:

Roles used for other members of household:

Sentence for Role "Witness:"	[W] is probably the male age [WM1] listed in the household of <[WM2],> [P1], in the 1800 Federal Census<of [L]><[WM3]><[M0]>
Sentence for Role "Witnesses:"	[W] is probably one of the [WM1] males age [WM2] listed in the household of <[WM3],>[P1], in the 1800 Federal Census<of [L]><[WM4]><[M0]>
Use:	These Roles are assigned in those rare cases when we can identify members of the household who have some other relationship to the head of household, for example a parent or grandchild. Enter the identified person, as a Witness. Assign the "Witness" Role when there is only one person listed in that age/sex bracket. Assign the "Witnesses" Role when there are two or more persons in the age/sex bracket, regardless of who the other person might be, or even if they are unidentified. If the "Witnesses" Role is used, enter the number of persons in that bracket in the first segment of the Witness Memo. Enter the age bracket in the first segment of the Witness Memo if the "Witness" Role is used, or the second segment if the "Witnesses" Role is used. Enter the relationship of the Head of Household to this person in the next segment of the Witness Memo. Examples: Witness:  age 10 to 15||his uncle Witnesses:  two||age 10 to 15||his uncle If there are any comment you want to add, place it in the last segment of the Witness Memo. This segment could also be used if one believes the spouse of a this person was in the household. In that case you might enter something like: and the male aged 60 to 70 is probably her husband, [R:Witness]
Notes:	The first Role above is the standard "Witness" Role with a custom sentence. Both Roles also have female sentences, with the word "male" changed to "female." If the need arises to record additional members of the household, one can add additional Roles, like "Witness2" and even "Witness3" with the same sentences.

Tip - Set Sex and Age Parameters for Each Role

On the lower right corner of the "Roles and Sentences" tab of the Tag Type Definition screen is a box that allows you tell TMG which sex, and the minimum and maximum age the role is appropriate for. I find it helpful to set those parameter for the wife, son, and daughter roles, as TMG then provides a warning if you inadvertently assign a role that does not fit the person to which it is assigned.

Entering Data in the Census Tag

• Enter the head of household as Principal.

• Enter the location, with as much detail as is available, in the Place fields.

• If the wife can be identified as probably being one of the females listed, enter her as a Witness, and assign the appropriate "wife" or 'wifes" role.

• If any children can be identified as probably being in the household, enter them as Witnesses, and assign the appropriate "son" "sons" "daughter" or "daughters" role.

• When the wife or child is listed in the census as one of 2 or more people in the same age/sex bracket, enter the number of people in that bracket in segment 1 of the Witness Memo for that person.

• In the rare case any other person can be identified as in the household, enter that person as a Witness, and assign the role of "Witness" or "Witnesses." Enter the age bracket, e.g. "age 10 to 15" in the first segment of the Witness Memo. If the "Witnesses" role is used, enter the number of people in that age bracket in the second segment of the witness memo. Optionally, enter the relationship of the person to the head of household in the next segment.

• The Memo field is the main tool for recording the relationships for the list of household members for the head of household narratives. See the following section for details.

Creating the Head of Household Narrative

The tag Memo field is used to construct the key part of the narrative for the head of household - both the recorded makeup of the household, and who we believe the persons counted actually were. Here is an example of what such a memo would be:

two males under age 10 (probably sons [RG:sons under 10]), one between 16 and 25 (unknown), and one over 45 ([PF]); and two females under 10 (daughters [RG:daughters under 10]), three between 10 and 15 (daughters [RG:daughters 10-15]), one between 16 and 25 (daughter [RG:daughter 16-25]), and one over 45 (his wife [R:wife over 45])

Note that the entry reports the number of people listed in each age/sex bracket, and then shows who we believe the people in that bracket to be. Since the sentence for the Principal introduces the Memo text with the phrase "with a household listed as" the entry in the memo starts with the first age bracket. Following each bracket, in parenthesis, is the person or persons we believe to have been enumerated, or the word "unknown" if I don't know.

The entry above would produce the following output for the head of household:

He appeared on the 1800 Federal Census of Pawlet, Rutland Co., Vermont, with a household of two males under age 10 (probably sons Joshua and Simeon), one between 16 and 25 (unknown), and one over 45 (Joshua); and two females under 10 (daughters Hannah and Welthyan), three between 10 and 15 (daughters Abiah, Polly and Drusilla), one between 16 and 25 (daughter Abigail), and one over 45 (his wife Hannah).

It would be possible to type the names into the memo instead of using the name variables you see here. I use the variables for four reasons: 1) it avoids mistakes typing the name, 2) the names automatically change if I later change the names of the persons involved, 3) when I output my data to a website using Second Site, each name becomes a link allowing the reader to jump to that person's entry, and 4) when there is more than one person with the Role, as with the children in this example, a single variable produces the entire list of all of them.

Should there be anything else of interest in the census, for example the presence of slaves in the household, I add that to the Memo at the end of the text listing the household members.

Troubleshooting Tips

Typing variables, especially Role variables such as used here, offers various opportunities to introduce errors. The problems don't appear until you create narrative reports. Here are some examples I've encountered:

In order to get the role names entered correctly, I find that it helps to enter the Witnesses before typing the memo. Then, I scroll the list of witnesses sot the desired person visible as I enter the Memo. That way I can compare what I type in the variable with the Role name for that person.

Roles Needed for Each Year

The age brackets used in the census differed from year to year. Some of the later years had a large number of brackets, creating the theoretical possibility of a huge number of Roles. But in practice, I have found that many of the possible cases do not actually occur. The list below shows those I have found sufficient, in addition to the roles for Principal and the all-purpose Witness Roles. Should a bracket not included be needed, that one role can easily be added without the need to add a complete set.

Copyright 2000-2008