Thursday, April 5, 2012

Those Pesky Penciled Census Codes in The 1940 Census


The Famous Ovaltine Decoder Ring
As genealogists, we strive to extract every last bit of information from a record. This, of course, is a good thing, all things considered.  

The problem comes when we try to read too much into a record and then start to see things that aren’t really there.   

Surely that mark or code must mean something important, we tell ourselves, otherwise why would it be there in the first place?

Do we need to get ourselves a secret decoder ring?

For example, now that the images of the 1940 Federal census are available for free in a number of locations online, people are flocking to the sites, downloading the images that they need, and now puzzling over some of the cryptic additions that were added to the census sheets once they reached the Bureau of the Census in Washington.

Could these cryptic codes added by the Census Bureau clerks contain additional information that would better help us understand the census entries in question?  If we knew exactly what they meant, might we better understand some small part of our ancestors’ lives?  Really, do we need that ring?

Actually… probably not.

Once the 1940 census sheets got to the Bureau of the Census in Washington, coding clerks went through them, looking for obvious mistakes, inconsistencies and so on.  An elaborate coding protocol was established so that the results could be quickly tabulated mechanically.  For example, if someone self-reported as “married”, but no spouse was present in the household, coding clerks were instructed to draw a line through the census taker’s “M” and write the code “7” in the space.  Thus a penciled “7” means “self-reporting as married, but no spouse present.” 

So, does this add anything to the researcher’s understanding of the census entry?

Actually… probably not.

The census entry clearly shows that the spouse is not present in the enumerated household.  It is not possible to tell from either the entry or from the coder’s actions whether the person coded as a “7” is married or not, even though he or she self-reports as such.

In reality, the coding clerks in Washington were not privy to any additional information.  Their only job was to code the information on the census sheets so that they could be machine-tabulated.  The coding operation was elaborate, employing, at its peak, 848 clerks working in two shifts.  There were general population coding clerks for the initial review, “comparison clerks” and even special editors to review and pass judgment on the new migration information contained in the question about residence in 1935.

Still, the information being coded was simply the information on the census sheets, as filled in by the census taker, nothing more and nothing less.  The Washington coding clerks were not endowed with special powers to know more about your ancestors than what was recorded by the census takers in the field, and even THAT information may be suspect.

Of course, sometimes, the code or correction was used to adjust a census taker’s failure to follow instructions.

For example, enumerators addicted to detail often put exact amounts in column 33 in spite of instructions to the contrary (column 33 was the “other income” column).  The Washington coding clerks were instructed to cancel out the amount and replace amounts over $50 with “yes” and amounts below $50 with “no.”  After all, that was what the question actually asked.  It was a “did you” question, not a “how much” question. Similarly, in the “income” column (Column 32) coders were instructed to cancel amounts over $6,000 and replace them with “$5,000+”

For those interested in the specifics of the Bureau of the Census’s 1940 coding operation, one of the best sources of information is an extensive document about the procedural history of the 1940 census on the IPUMS-USA website.  IPUMS is not a genealogy site; it is the acronym for the “Integrated Public Use Microdata Series” and is part of the University of Minnesota’s PopulationCenter.   Data geeks, statisticians and epidemiologists – all of whom use census data – are the primary users of this site. 

Here’s the specific link to the procedural history:  http://usa.ipums.org/usa/resources/voliii/enumproc1940.pdf

(Note: if you were at my talk at the wildly successful New York Genealogical and Biographical SocietyNew York Public Library’s “Road to the 1940 Census” event on March 24th, you would have found a link to the IPUMS site in my handout and would possibly have remembered my exhortation to the audience of nearly 500 genealogists to “Read the Enumerator’s Instructions and also Read The Procedural Histories”, both of which can be found on the IPUMS website.)

Bottom line:  Sometimes, what you see on the population schedule is actually all there is.  Coding clerks are rarely in a position to add new or more conclusive information.

Sometimes – and with all due apologies to Sigmund Freud - , a cigar is just a cigar, no matter how much we’d like to have it convey much more meaning.

4 comments:

  1. Thank you so much for this, very helpful!

    ReplyDelete
  2. Just found your blog via your interview with Gini at GeneaBloggers. I love what you are doing here.

    ReplyDelete
  3. Sharon pointed me to your post and your example is exactly the situation that's puzzling me! Thank you! I can now continue with my indexing. :-)

    ReplyDelete