What's the best way to enter census information? What's the best way to enter citation sources?
What's the best way to phrase the census sentence? The debate about the best way to deal with
census information in TMG was raging before the Maryland TMG User Group was founded in 1997 and it
still goes on. Although I know I've said it before, the best way is the way that works for you. Determine
what you want your census evidence to provide and make TMG give it to you.
To decide what works best for you, answer these questions. The answers may help you consciously
determine exactly what you want from your census data and, hence, what you expect TMG to do.
The census is a bit of evidence in your research. How do you analyze it?
The census is your research skeleton. How do you track your research progress?
Do you return frequently to the original census image or transcription to check on some fact?
When reading journal articles, what census information are you happy to find in the footnote?
When reading journal articles, what census presentations impress you most?
My TMG "census procedures" will be presented here. Not only will you see what I do, but why I have
chosen to do it. My system works for me and it might give you some ideas. For additional points of
view, investigate these TMG user pages:
An isolated census is not worth much. It may provide some evidence, but its true value appears when
placed in context with other censuses and contemporary documents. "Benjamin Gifford appeared as head of
household, age 21, shoemaker, born in Pennsylvania, in the 1850 census of Tioga Co., Pennsylvania," doesn't
say much. I want a report that lets me see all the census data about a person, family, region, or time
period at once. TMG will not print up reports that resemble some of the excellent
charts and diagrams found in the National Genealogical Society Quarterly, but these reports help me
analyze the census evidence, see patterns, give me ideas for further research, and help me create graphic
analyses more easily.
A report that shows what census information, both positive and negative, I have on a research group
A spreadsheet report showing all members of a research group and their residences each census year to
allow me to chart family migration patterns
If you're interested in any of these reports, follow the links to see how they're created. Some of these
reports are dependent on data input. For example, you can't create a list of households in dwelling and
family number order unless those numbers are entered in a data field.
Ah, the great Census Sentence. How should it read? The introduction of Roles in TMG
turned my simple census sentence on its head. Like many users, I experimented with various possibilities,
but no matter how I reworked TMG's sentence variables, the end result still didn't come close to my benchmark,
articles in the National Genealogical Society Quarterly. I returned to basics.
Census Tag sentence structure for 1850 Census and later:
-[:CR:][:TAB:][P], age x, appears as head of household in the census <of [D]> <in [L]>< with [WO]>
-[:CR:][:TAB:]<[M0]>[W] appears as Mary Smith, age x in the census <of [D]> < [L], >in the household of [P]
as his [relationship stated in census]
Both sentences are similar to those used in the Sample data base. Words in italics are added as necessary
to state precise information found in each census.
Cens-stat Tag sentence structure for statistical censuses (user-created event tag):
-[:CR:][:TAB:][P] was head of household on the census <of [D]> <in [L]>
-[:CR:][:TAB:]<[M0]>[W] is probably the male/female age x to y enumerated in the household of [P] <in [L]> <[D]>
There are several items to note in these sentences.
[M0] (that's M-zero) is a wonderful little trick that keeps a memo from printing. I transcribe the
census in the memo field. It will print as a footnote or endnote just after the citation for the census
event. It will not print when the witness sentence appears.
[WO] (that's W-capital O) lists all other people named in the census. Because TMG 5.x allows
witnesses to be entered with name variations, the name used will be that found in the census.
Finally, note that I began both sentences with a hyphen. I have excluded both sentences. They will
not be printed, unless I check "Show excluded data" in the Report Options. In other words, I don't
use census sentences in my narratives at all! Instead, I took a page from my friend who writes out
his narratives in a formatted note field. I try my best to weave the relevant census information into
my ancestors' story or set aside a section for census analysis. For an example of the former, read the
Benjamin S. Gifford narrative.
By the way, this is an example of a TMG Journal Report in HTML format.
Why do I enter my census information this way? It's a result of my answers to some of those previously
First, I found that I did return frequently to the census transcription. I needed to see the family
as a unit. In many cases, I needed to see the family on the census page among the other families in the
neighborhood. If I treated the census entry as simply bits of evidence as to name, birth date, and birth
place, I lost the ability to easily view its context and the important evidence that context might provide.
Therefore, I transcribe the census in the memo field. With all witnesses attached to the census event, this
memo is easily accessed from any witness page.
In those thankfully rare instances where I want to look at an entire census page, I simply create a
note event with enumeration date and place, attach a digitized image of the page to the event, and link all
names on the page as witnesses. I don't print out my note events in any report. They exist to make
my research ideas easily available to me.
I don't like canned narrative reports. Even TMG's flexibility still produces reports that are stiff and
repetitive. You will never see an article in the NGSQ that repeats "John Smith appeared as head of
household . . ." over and over. Therefore, I chose to write out my reports by hand and keep them in what
I termed a Comment field. These fields can be formatted as if in a word processor, but can be easily
edited when I work on the database. For me, this combines the flexibility of the database with the finishing
techniques of the word processor.
I still need the Census event tag, though, to produce the reports I use to analyze census evidence. I
simply left the simple sentence structures in case I wanted a research report that used their content.
Census Tag Entry Screen
Another point of debate arises occasionally. Should you use the official enumeration date of the census
or the date the census taker visited the household for the Census event date? Remember, all questions were
supposed to be answered relative to the official enumeration date.
I enter the census taker's date. I can always look up the official date if I need it, but the census
taker's date is not easily accessible if I don't record it. Although the official date was supposed to
regulate answers to census questions, that was not always the case. It's nice to have both dates available
when analyzing the information.
Finally, many TMG users create a new census event tag for every census year: 1790-CENS, 1800-CENS, etc. For me,
that would mean 43 separate census event tags at the moment. I'm a lazy person, so I like to keep things
as simple as possible. I can see two possible reasons for creating more than one census event tag:
Do you want to create a different sentence or different roles for each census year? I did want a
different sentence for any census that didn't name all household residents, hence my two census event tags.
Do you have a special report that can only be generated if each census year has a different tag? I
haven't thought of one, but one might well exist.
There is one major debating point in the question of how to document a census: what constitutes the
A geographic entity such as county or city?
A physical entity such as the NARA microfilm roll or the census volume?
Source Definition Screen: General Tab
There are pros and cons for all three possibilities. I use the physical entity, microfilm roll, volume,
or CD, as my basic source unit. It results in far fewer entries in the Master Source List than entry number
3 and requires less content in the Citation Detail than entry number 1.
I am a 'minimal user' of the Split Citation Detail, using only CD1 and CD2. I decided that I would need
to closely edit any report I submitted for publication, so the flexibility provided by nine split citation
details was moot. As long as the citation contains all information necessary to identify and locate my
source and determine its reliability, the exact form was immaterial.
Terry Reigel's documentation method is similar to
mine and he makes extensive use of the Split Citation Detail.
I frequently include evidence analysis points in my Citation Detail field. For example, if the 1850
census is my source for a circa 1813 birth year, I include "age 37" as part of my citation. After all,
that is what the census stated. A birth year of "circa 1813" is my interpretation of the data.
Source Definition Screen: Supplemental Tab
I reserve the <comments> field for two pieces of information: problems with the actual census, such
as legibility, missing pages or townships, or drastic date variations, such as an 1810 census being
taken in 1812; and background information derived from the census itself, such as
population count or mean and median property values.
Another frequent question is, "What is the right source abbreviation for a census? Should I
begin the abbreviation with CEN? with Year? with Geographic location? with Head of household?"
The right source abbreviation is the one that lets you find your source easily. For a census, I use
the following: County Co ST Year (microfilm number), e.g., Bradford Co PA 1850 (756).
When I began my research, censuses were found on microfilm, in original volumes, or
transcribed. Now, digitized images are available on CD-ROMs and online. Should you add a separate source
for each of these media?
Theoretically, the NARA microfilm and the digitized images are the same, but the latter have been
"enhanced" for better legibility. This enhancement does make them easier to read, but I have several
instances in which the enhancement has done its job too well. For example, an age of 17 on the microfilm
copy has become age 11 on the digitized image. I enter a new source if I'm using a different medium.