"Over Count" of Chinese in the 1940 Census?

The U. S. Census is an invaluable resource for historical and genealogical research. I was excited when the 1940 U. S. Census was released earlier this year, but frustrated because it was released without a name index, which made it virtually impossible to use and maintain sanity. Fortunately, it is now much more functional with the addition of a name index generated with the help of hundreds of volunteers who transcribed information from the original handwritten record sheets collected in the field to create a searchable online digitized database.

I gave it a test run by first trying to locate the record of our family in Macon, Georgia.

That test was not a difficult search since we were the only Chinese in the entire city at the time. This initial success led me to make several additional searches which, much to my surprise, suggested that the U. S. Census may have over counted the number of Chinese in many states. Before I explain why I reached this conclusion, I first need to retrace the process to show how I discovered the errors. So, be patient!

How to Use Ancestry.com to Search the Census Database

On the initial search screen shown below, you enter as many of the listed fields as possible: first and last name, birth year, exact or within a range, location (where they were born and location and approximate year of “any event” (marriage, divorce, for example).

Choose “match all terms exactly” to prevent finding too many cases, but if that fails, you can then widen the yield by removing that constraint.

Depending on how unique your target is, you might locate that person by entering only part of the requested information.


On lines 42-47 of the census record shown below, our family is listed for 1940 at 519 Mulberry Street. I highlighted in bright green how our race was coded, "C 4."

Failure Locating My Uncle

I also wanted to find my uncle who came to Atlanta in 1937, but I had no success finding him. Possibly, he was not in his laundry on the days when the census taker came. Or, I used the wrong spelling of his name or the census taker, or transcriber, entered it differently from the way I spelled it.

If you examine a sample of original record sheets, you can easily understand why these errors occur because some of the handwriting is poor or the record sheet did not scan clearly.

An alternative approach is to use Advanced Search, which allows you to specify several filters such as GENDER, and more importantly for my purpose, Race/Ethnicity. By adding CHINESE to the Race/Ethnicity field (see bright green field) I was able to retrieve only the Chinese cases for a specific Location (you can specific a City, State, or Both). I specified Georgia (See bright green field).

 Since there were very few Chinese in Atlanta in 1940, my thought was that my uncle’s name would come up even if they misspelled his name. Then since I knew the street address where he lived, I should have been able to locate him even if his name were misspelled. Of course, this method would not work easily for cities where there were hundreds of Chinese.
(My brother George was listed as "Georgia" and my sister Eugenia as "Eugennia," and Joe Yip in Mississippi was listed as "Qck Zip")

However, even this approach failed to locate my uncle in Atlanta in the 1940 census. He must have been out of his laundry when the census taker came.

Searching For Trends or Totals

Instead of trying to find a specific person by name, one might be interested in seeing patterns or trends in the number of Chinese over decades. Thus, I looked at changes in the total number of Chinese in Georgia for each decennial census.

Results for Chinese Living in Georgia

When I searched for CHINESE in Georgia for all decennial Censuses, I found a gradual increase from 203 in 1900 to 445 in 1940.

I then restricted my search to Chinese in Georgia who were born in China. The percentage declined with successive decades. Only about half, or 222, of the Chinese living in Georgia in 1940 were born in China. In contrast, a decade earlier in 1930, around 80% or 206 of the 253 Chinese in Georgia were born in China and in 1900, 100% of the 203 Chinese in Georgia were born in China.

Some Strange Findings

When I examined some of the cases classified as Chinese on the index lists, many of their surnames, e.g., ADAMS, did not seem to be Chinese. Nor were their occupations typical of Chinese immigrants such as laundryman, restaurant cook. or waiter.

I therefore examined images of the original census record sheets for these suspicious cases.

Below is a 1940 census record for one Adams family in Missouri that was indexed as Chinese.

The original handwritten census record appeared with the searched name highlighted in yellow, and family members in green. Beneath this info shows how it was entered in the searchable database.

On the original sheet for the Adams family, RACE had a code of “C 2“ but they were recorded in the Index as “Chinese.”[See the Column I highlighted in bright lime green] In other words, in the 1940 Census many “C 2” cases that should have been recorded as Colored or Black were counted as Chinese.

The same error occurred in other states I sampled including Missouri, Mississippi, Delaware, North Carolina, Nebraska, and Alabama. This error did not show up in Minnesota, at least for the the cases I sampled!

Other support for this conclusion is that the RACE code on the original sheet (See the first image on this post) for my own family was not "C 2," but "C 4." [Actually, it seems to have been “Chi” but that was lightly crossed out and replaced with ”C 4.”

As a type of 'reverse' check, for RACE/ETHNICITY I searched for COLORED (the term in 1940 for Blacks). When I examined the original record sheet for cases retrieved in this search, the code was usually “C 2.” [In one state it was “Col” and in another, “Neg.”]

Birth Rates for Chinese in Georgia

The misclassification error also occurred for birth data.
I did a a search for Chinese births in 1939 in Georgia on the 1940 census.

Given that the majority of Chinese in Georgia, as elsewhere, were single men, or men with wives in China, one would not expect them to have had many babies born in Georgia.

[Therefore to get a fair-sized sample, I used a range of years from 1929-1939] In this decade, there were 56 Chinese females born (and roughly the same number of males).

[Note: the point of this analysis is not to determine the number of Chinese babies but to examine the accuracy of the coding.]

A look at the Table below shows many of the surnames of the babies did not "look like" Chinese surnames.

You get definitive evidence of errors if you examine a sample of the original record sheets for births with a non-Chinese looking surname. The RACE column for these cases shows a code of “C-2” or Colored (the prevalent term for Blacks in that era) as in the record shown below for Bessie Butler and her relatives. In other words, these babies got misclassified in the index as Chinese, for whom the correct code is “C-4.”

While one hopes for, but does not expect 100 percent accuracy for such a complicated undertaking as taking a national census, egregious errors such as the ones described above simply should not have happened and could have been caught with minimal review.

1 comment:

  1. Hi,

    had to come over here and review this post as well. This really helps and I'm so glad I found this post. I was starting to wonder if I had gotten my family history all wrong and we were indeed mixed with Chinese and not Native American after all.

    This clears up a lot although, I still have a few questions.

    Thanks for all you do.




About Me