This is actually an excellent illustration of how organizational incentives seep into data.
Imagine you have a historical temperature series record. You have an bureaucratic incentive to find and justify an argument as to why past temperatures should be adjusted downward, making the curve of global warming look steeper. You have no bureaucratic incentive to look for the opposite.
The result is that, overall, modern corrections adjust past temperatures downward -- what we'd expect if these biases exist. But this doesn't mean any specific correction is based on "fabricated" research. It's probably perfectly valid. It's just that downward corrections are more attractive to research, because they produce research that has a social impact that the field finds desirable.
Imagine you have a historical temperature series record. You have an bureaucratic incentive to find and justify an argument as to why past temperatures should be adjusted downward, making the curve of global warming look steeper. You have no bureaucratic incentive to look for the opposite.
The result is that, overall, modern corrections adjust past temperatures downward -- what we'd expect if these biases exist. But this doesn't mean any specific correction is based on "fabricated" research. It's probably perfectly valid. It's just that downward corrections are more attractive to research, because they produce research that has a social impact that the field finds desirable.
(Edit: here's a picture of the net adjustments: https://judithcurry.com/2014/07/07/understanding-adjustments...)