How reliable is data for slaves enumerated in the 1830 complete sample?

Under Known Issues, it reads “Use with Caution.” What does this mean? I found some unusual observations – which may be coding errors. Because the names of the household heads have been removed the record, I can’t compare the record with the original manuscript census.

So, how confident should I be with these data?

Here is a link to a page listing some of the known errors with the 1790-1840 datasets. In particular relation to the enumeration of slaves it looks like the counts are not consistent with other published figures and are thus not representative at the national level.