“Trust, but Verify” was an oxymoronic slogan from the era of nuclear weapons treaties during the Cold War. Diligent researchers understand the value of that approach. Restated and applied to genealogy, the rule is, “Never dismiss out of hand any documentary evidence, including census data, but don’t assume census data is always 100% accurate.”
There is clear rationale for that caution. First, census data is subject to error multiple times. The person supplying information to the census taker can be mistaken as to any number of things such as ages or places of birth of people in the household. The census taker can record the data incorrectly. Further, the data collected was organized and rewritten into a final document. Each reproduction of the census information presented an opportunity to introduce errors, including misreading another person’s handwriting.
In addition, census data was not subject to the same checks and balances as other official documents. For example, original deeds and wills copied into court records benefitted from court oversight of the process. Witnesses attested to the accuracy of those document, heirs could question a misstatement that affected their interests, and neighbors could request resurveys of land boundaries they thought to be in error. No such process accompanied the tabulation and publication of a census. As a result, that data is far more prone to error than other records.
I recently ran across two illustrative errors in the same census entry. Searching for Henry Willis, carpenter of Maryland and Philadelphia (1829-1906), I found the family of John and Rebecca Kilgore Willis of Cecil County, Maryland. They had six sons and four daughters. By 1850, three sons were of age and no longer listed in John’s household. Hoping one of the three might be Henry, I looked for them in the 1850 census. I did not find Henry, but found James Willis and “David” T. Willis living next door to each other. The census entry showed the following:
123 Sarah H. Shivery 27 M
James Willis 21 F
Mary 3 M
Joseph 1 F
Sarah A. 27 M
124 David T. Willis 22 F
Hannah A. 3 M
George A. 1 F
Margaret R. 5/12 F
Mary E. 40 F
125 Hannah Terry 9 M
The problem with this data is obvious: the genders and ages do not match the named people. Whoever completed the census form moved that information up one line from its proper position. James Willis’s proper age and gender are 27 and M. That data is shown on the form one line above his name. It is incorrectly associated with a child named Sarah H. Shivery who is the youngest daughter of George Shivery in the adjacent family #122.
No problem. To get the correct information, just mentally move the data down one name.
However, that is not the only error. James Willis’s neighbor is supposedly David T. Willis with a wife Hannah A. Willis and several children. However, Daniel Willis, not David, married Hannah Ann Sutton on 15 April 1847. In fact, there was no David Willis in that location in 1850. Whoever entered the data in the census form apparently misread someone else’s handwriting and thought the name Daniel was David. That is not hard to do. A script “n” can easily be mistaken for a “v” and the “el” as a “d.” Try it in your own handwriting to see how easy it is to make the two names look the same.
Of course, the opposite could be true, Maybe it was David Willis who married Hannah Sutton, and there is no Daniel Willis. This is where the “Verify” part of the slogan becomes important. The proof is found in subsequent records. Daniel Willis registered for the civil war draft in 1863 and appeared in the 1870, 1900, and 1910 censuses. David did not – – because, of course, he did not exist.
So, the message is to confirm the data found in censuses with other sources. Many people on Ancestry.com have not done so. As a result, there is a fictitious David T. Willis running amok on many trees. We all make mistakes. They come with the territory. This is a zillion piece puzzle, and we only have a few thousand pieces available to make sense of the picture. However, diligence can easily eliminate some errors. It is worth the effort.
 1850 Census, Cecil County, Maryland https://www.familysearch.org/ark:/61903/3:1:S3HY-XCHQ-7X5?i=109&cc=1401638&personaUrl=%2Fark%3A%2F61903%2F1%3A1%3AMD46-PD2
 Cecil County Marriage Licenses 1840-1863, Genealogical Society of Cecil County, August 1990, 20 at https://web.archive.org/web/20150214151843/http://www.cecilhistory.org/virtuallibrary/marriage3.pdf
 Civil War Draft Registration 1863-1865. https://www.ancestry.com/discoveryui-content/view/3912656:1666?tid=&pid=&queryId=66dc11368cf6a04f330078b4413841ae&_phsrc=Uwh1&_phstart=successSource
 1870 Census, Cecil County, Maryland. https://www.familysearch.org/ark:/61903/3:1:S3HY-D15S-Y4W?i=4&cc=1438024&personaUrl=%2Fark%3A%2F61903%2F1%3A1%3AMN38-D3L
 1900 Census, Cecil County, Maryland. https://www.ancestry.com/discoveryui-content/view/20341258:7602?tid=&pid=&queryId=66dc11368cf6a04f330078b4413841ae&_phsrc=Uwh3&_phstart=successSource
 1910 Census, Cecil County, Maryland. https://www.ancestry.com/discoveryui-content/view/10469167:7884?tid=&pid=&queryId=66dc11368cf6a04f330078b4413841ae&_phsrc=Uwh2&_phstart=successSource