A very large number of humans confuse noise with the real signal.
iFor economists-stuff like inflation, gdp, jobless rate, etc. are all numbers which have a lot of noise in them. Trying to pinpoint whether the inflation rate is 2% or 4% by measuring a basket of goods is a waste of time. In general, when there is money around in a country-people know it-prices go up everywhere, of everything, and then it is easy to see high inflation. In other words, something like 8 to 10% inflation is distinguishable from 2-4% inflation, but trying to get it to a decimal between 2 to 4% is really not right.
Think of the elephant and blind men story. A set of blind men who were touching different parts of the elephant had a completely different explanation of the animal or thing they were touching. The guy who touched the trunk had one version, the guy who touched the tail another, etc. Nobody could "see" the big picture, the complete elephant.
This is where a lot of data taking makes a huge error. Social sciences are full of all these random observers-people like Levine correlating crime rates with single parent homes, or others trying to correlate daily chocolate intake with happiness. It is good to sometimes first try to see the big picture before trying to take minute data. Your conclusions based on your data may not have anything to do with reality.
A lot of things which happen around us are full of random noise. Trying to mathematically smooth out the curve by taking more data. etc. is not gonig to change how reality functions.
Economists, weather and environment scientists, social scientists, etc. are the most liable to make these errors, because they are observing very complicated phenomena. Medicine was like until 1830 or so, before it became more organized and "scientific" and we could correlate causes with effects better, see the real signal and not just draw consclusions based on the noise.