Monday, August 17, 2009

Searching the 1930 US Census on

I realized over the weekend that I had not checked the 1930 U.S. census on for my Thompson brick wall problem. I also realized that when Footnote brought the 1930 census online that I used it only briefly, in connection with the Footnote Pages and the I Remember application on Facebook.

Here is a quick look at a typical search in the 1930 US Census on, which is FREE to search until 31 August.

I searched for Lillian (Daniel) Thompson, wife of Robert Leroy Thompson. Lillian was born in 1900 in Tennessee, but I wasn't sure where the family lived in 1930 - it might be Tennessee or surrounding states. I decided to start with wild cards in the first name and last name - so I chose "Lil*" and "Thomp*. Here is my search sequence:

1) I started on the 1930 Census page on Footnote at and entered "Lil*" and "Thomp*" in the search fields:

2) The Matches screen showed that there were 2,374 matches for my search criteria (two screens below):

That is way too many for me to search page-by-page, so I decided to narrow the search. The search can be narrowed by clicking on one of the items in the left column - they are Last Name, First Name, Place, Year, Age, Browse Description, County, Enumeration District, Estimated Birth Year, Family Number, Sheet Number and State.

3) I chose to narrow the search by Birth Year, so I clicked on that:

A list appears just to the right of the left-hand column with the list of entries.

4) I scrolled down and chose 1900:

The list of matches was pared to only 54.

5) I decided to narrow the search further and clicked on "State":

This gave me the list of States with a Lil* Thomp* with an estimated birth year of 1900.

6) I chose Virginia:

There were only four matches for Lil* Thomp*, 1900 birth year and Virginia. Unfortunately, none of them were the one I want. And none of the other 2,370 persons from the name matches were the one I wanted.

I know that I've been spoiled by the search field options for the census records. By comparison, the search options are primitive and work much slower than the equivalent searches.

In my opinion, really dropped the ball when they indexed the 1930 census. I don't know who indexed it (was it FamilySearch teams at NARA?) but they should have included more fields. If they were limited to a certain number of fields, they really should have chosen different fields.

For example, in my humble opinion:

* What value is there to the "Family Number" and "Sheet Number" entries? The user can narrow by "State," "County" and "Browse Description" (like City or Township) - that should be sufficient. [See Update below!]

* The user can narrow the Search by using only one parameter at a time. This is extremely time consuming and frustrating to an user. My preference is a set of Search fields that include all of the indexed fields so that I can enter my name, birth, location, etc. information once and do a narrowed search.

* The 1930 US Census day was 1 April 1930, and the age as of that date was enumerated. To narrow by age, the researcher has to choose a specific "Age" or "Estimated Birth Year." A person born in 1900 would have their age listed as 29 or 30. There is no option for an age or year range in order to encompass small errors or birth date relative to the census date. At a minimum, there should be a plus or minus 2 year range built into the search just so the user doesn't have to go back and enter a different year several times.

* There is no search capability on Birth Place. This is one of the most critical search fields for me. I am amazed that this field is not available to use to narrow a search!

* There is no search capability for other family members - spouse, parents (one or two) or children names. This is one of the quickest ways to narrow a search for a common surname.

* There is no search capability by sex, race or relationship to head of household. These can be very useful to narrow a search.'s 1930 census index has several flaws also - I discussed these in Quirk - 1930 Census Index.

So we have the reality that there are three different and flawed 1930 census indexes available (note that the HeritageQuestOnline 1930 census index is only a partial index). We are stuck with this situation - I doubt that any company will go back and re-index the census. So we need to understand what each index does and doesn't do, and work with it.

Why even use more than one 1930 census index? Each can be used, but none is as good as it can be. However, there have been at least three different indexes created by different parties, and we have no guarantee that any two of them used the exact same census microfilm pages. There is the possibility that one index is more complete than another. Therefore, a diligent researcher will search all of the available census indexes in their quest to find their elusive ancestor.
UPDATED 11:30 a.m.: Cathi Desmarais on Facebook suggested that the Sheet Number and Family Number are indexed because they are used in the source citation of the image. I see the Sheet Number in the Source Citation, but not the Family Number. I didn't even think about this possibility when I made the comment. Note to self: think outside the box once in awhile! Thanks Cathi - a gold star for you!


Elyse said...

Great post Randy! You made so many great points. I hope you end up breaking down that brick wall soon!

Lynn said...

I have to agree with what you said about, Randy. I might add that I tried the 1930 census today, and found only 4 of the families I was looking for - and that by creative searching. One was okay, but three had first names that were badly transcribed. The names should have been transcribed correctly, as they were written correctly by the census taker. But they had "Kunh" for Knute, "Flans" for Claus, and "Ada" for Ida. Fortunately the surnames were correct and I knew where all these families were living at the time, or I may not have found them.