But even "raw data" is astoundingly biased.
How do you think that "raw data" was collected?
For an example, let's take the U.S. Census. Goal: count the number of people living in the U.S. (Tied in, of course, with the political reasons for the census, which dictate who is counted and how their categorized.) We could go on for a long time talking about the methodology, but we'll skip just to the part people argue about: whether to use the "raw data" or the statistically corrected data.
The census always undercounts people, but it does so unevenly. Urban areas have a much higher percentage of undercount than suburban and rural. However, it's possible to recalculate the census numbers using statistical sampling to get a more accurate, but less raw, count.
So which is the unbiased data? Guess what, there ain't any.
(See Census Sampling Confusion, The Fiscal Impact of the Census Undercount on Cities.
Even in physics, it's impossible to get unbiased information, as the act of measurement changes the measured system. Fortunately, the change can often be less than the precision of the measurement.
And more importantly, the bias comes in first in what we choose to measure.
[ Parent ]