There is no question that mistakes in the censuses abound, particularly in the early years. For example, following the 1840 census, analysts noted that the data showed free blacks were ten times more likely to suffer from insanity than slaves. Moreover, rates of black insanity increased the farther north one proceeded. Pro-slavery advocates seized on this data to show that slavery was a blessing for blacks: African-Americans, they charged, were obviously mentally incapable of handling the pressures of freedom. Others, however, combed the returns for errors. Cross-checking variables, they found high rates of black insanity in many Northern towns which contained no black population! Scholars today believe that the data is badly flawed. As Patricia Cline Cohen has explained in A Calculating People: The Spread of Numeracy in Early America, the unwieldy, eighty-column schedule made it easy for enumerators to enter answers in the wrong column. In particular, the column headings made it easy to record "insane or idiotic" whites in the "colored" column and, indeed, in most cases, the insane blacks were listed in all-white households.

Other data errors stemmed from poorly worded questions.

Furthermore, many enumerators were creative spellers, mangling unfamiliar surnames.