If you get rid of the political and social impact of AGW and just look at the problem they have as demonstrated by HARRY_READ_ME it becomes pretty clear. They have data from land stations that move over time, stations that don't report consistently, C02 data, data from tree rings in the last 20 years that breaks down, cloud and precipitation data that doesn't match the land station data values or locations, a cooling trend recently and a warm period 500 years ago all from different people at different times in different formats with probably different goals and different agendas. And they suspect that the earth is warming and want to perform a robust analysis of CO2 versus temperature across the surface of the earth. Compound it all with forcing effects of volcanism and latitude. They need better algorithms and programmers to have any real trust in the output they are getting yet they press on and just force the data to reveal their "truth". AGW in this context is not a cohesive theory supported by all the evidence. That's bad mojo if you need the next GW grant to eat. Their solution? Fraud. The data is insufficient or contradictory and it should be dropped, but instead they use synthetic data when "necessary".
GIGO (Garbage In, Garbage Out) sums this whole deal up.
(Okay, fraud on a truly grand scale should be included, but it screws up the acronym.)