One of the really interesting things in Bret Devereaux's post on Tiberius Gracchus is how modern historians, through archaeology and modelling, have a better understanding of the economic and demographic situation in 2nd century BCE Italy than Gracchus did from the census data. They were able to collect evidence of land use changes and construction taking place, and analyse that data, to show there was a growing population of smallhold farmers. Meanwhile Gracchus was reliant on a self-reported census, which (according to the argument) was declining because people lied to avoid being eligible for conscription.
The Romans didn't really have the resources to go out and audit everyone, or do a separate land-use survey to check the census against, but over 2000 years later we were able to go back and show discrepancies between census returns and actual land use. It's fascinating!
