I have tried to compare the 1900 full count data within state and county to the U.S Census bulletin for the decennial census. I found the numbers not to match. Using Alabama as an example, the number of full count for each county varies substantially from the Published census numbers. Is there a reason for this?
The Census Bulletin from the 1900 census includes population estimates by state and county for the year 1900. These estimates are very close to the population estimates you can obtain using IPUMS USA microdata from the 1900 full count census; however, as you noted, they are not identical. The full count census data provided by IPUMS USA has been transcribed by Ancestry.com and edited and harmonized by IPUMS. In the process of providing these historical data, it is inevitable that some records are lost or incomplete, making them unusable. There can also be some instances of duplicate records appearing in our data, which we work to correct. For these reasons, the exact number of person records retained in the data we provide will not exactly match the population estimates derived by the Census Bureau in the year 1900. If you see any discrepancies that are particularly large or if you have concerns that are not covered here, please feel free to follow up.