“Wait a minute,” you’re probably saying. There is no such thing as a 1917 US Census. There was a 1910 and a 1920, but no 1917. Well, you’re right, but there was in the Virgin Islands. The US took possession of the islands on March 31, 1917 and placed the islands under control of the US Navy. The Government decided to take an official census of all residents of St Croix, St Thomas, and St John as of November 1, 1917. When the 1920 census rolled around, they figured that they had just done the Virgin Islands, so they included the 1917 census as part of the official 1920 US census. If you look on census websites for the 1920 census, you will find that the pages for the islands are actually dated earlier. Of course, being the islands, the census for November 1 wasn’t actually begun until after Christmas, and wasn’t complete until almost February. Island Time isn’t new!
The 1917 census is simpler than the 1910 US Federal census, asking for 21 items per person rather than the 32 items in the 1910 US census and the 31 items in the 1910 Puerto Rico census. In particular, while the US census asked for places of birth of each person, their father and mother, the VI census only asked about the person. The census asked about citizenship, but the categories were “Virgin Island”, “Danish”, and “Not Declared”. According to the treaty ceding the islands to the US, each citizen was allowed to choose whether to remain Danish citizens or to declare their citizenship of the Virgin Islands. These did not make them US citizens, however. The VI citizens did not get automatic US citizenship until 1925, however some people naturalized in the interim.
The US census was interested in tracking unemployment, so it asked about employment in 1909 and 1910. No such questions were on the 1917 census.
There are a couple questions on the 1910 census that are very helpful to genealogists that were, sadly, left off of the 1917 VI census. The US census asked for the “Number of years of present marriage” which pins down marriage dates, and it asked for numbers of children born and number living, which is very helpful in knowing if we have found all children born before 1910. All of these questions are missing from the 1917 census.
Another interesting difference speaks to the cultural differences in DWI society and US society. The earlier DWI censuses always asked for Religion. The US census never did. The US census asks for Race or Color. The DWI census never did.
For all the shortcomings of an abbreviated special census, the 1917 Virgin Islands census has one very significant advantage over the 1910 US census. Since it was a special census, the US Census Bureau assembled a detailed analysis and published it on June 15, 1918. The report gives 200 pages of statistics and analysis on the islands. Much larger than the anemic 25 pages or so in the census reports since. The report discusses the data collection, society, geography, and many pertinent statistics. It is an outstanding reference. The report is available as a free download at www.census.gov/history/pdf/1917usvi.pdf. This is an excellent read and highly recommended. (disclaimer: I was a Statistics professor for many years, so we may have different definitions on “excellent read”)
Of particular interest and relevance to some of my earlier posts, the report describes the data on marriage. The 1910 US Federal census recognized 4 possible marriage statuses: Single (S), Married (M), Divorced (D), Widow(er) (W). Starting in the 1910 Puerto Rico census and continuing in the Virgin Islands censuses, they added a 5th category: Married Consensual (MC). This term is used for people living together without being married, or what later became known as “Common Law Marriage”. This was so widespread that the report gives a discussion of it, together with a discussion that the statistics are likely understated since many MC families self-reported their status as “Married”.By 1930 the MC category was adjusted to mean “Mutual Consent”, but otherwise it meant the same thing.
Beginning in 1930, the Virgin Islands participated in the US Federal Census and the forms were the same as Puerto Rico (although the PR forms were in Spanish) and much closer to the US Federal form. The 1940 census will be released in 67 days, which will bring the census coverage for the Islands to 99 years (1841-1940).
If you find my discussion of available records for the islands interesting and/or helpful, post a comment to let me know and I’ll be happy to post similar discussions of the other records I have found.