I have Bolivian census data for 2001 and 2012. There are noted discrepancies between the two, particularly the significant drop (as % of total) for "indigenous" population between the two. However, the data for 2012 seems to be on the basis of the total, but the 2001 data seems to be only on the basis of those who answered the question (in 2001 "non-indigenous" was an option, in 2012 it was not). Taken together, this suggests that the 2001 data significantly *overcounted* the share of indigenous population. However, I'm not a demographer. Any advice on how to proceed with using this data?
The discrepancy is large. If only looking at those who answered the question, then in 2001 the indigenous population was about 62% and dropped to about 40% in 2012 (which was calculated using the "didn't answer" figure). But if we use the "no answer" data from 2001, then the indigenous population in Bolivia dropped to 38% (a number closer to the 2012 figure). But does this mean that everyone has been overestimating the share of indigenous population for Bolivia for the past decade?