Background
Nearly everyone who claims that the rate of global warming experienced over the past century was unprecedented over the past one to two millennia also claims that the data used to support that claim have been appropriately adjusted for the urban heat island effect and other non-climatic phenomena or that such non-climatic phenomena only trivially perturb the pure climatic background signal contained within the data.

What was done
In a critical examination of these multiple claims, McKitrick and Michaels (M&M) calculated 1979-2000 linear trends of monthly mean near-surface air temperature for 218 stations in 93 countries, based upon data they obtained from the Goddard Institute of Space Studies (GISS), after which they regressed the results against indicators of local economic activity -- such as income, gross domestic product growth rates, and coal use -- to see if there was any evidence of these socioeconomic factors impacting the supposedly "pristine as possible" temperature data. Then, they repeated the process using the gridded surface air temperature data of the Intergovernmental Panel on Climate Change (IPCC).

What was learned
M&M report that the spatial pattern of trends they derived from the GISS data was "significantly correlated with non-climatic factors, including economic activity and sociopolitical characteristics."  Likewise, with respect to the IPCC data, they say that "very similar correlations appear, despite previous attempts to remove non-climatic effects."  These "socioeconomic effects," in the words of M&M, "add up to a net warming bias," although they state that "precise estimation of its magnitude will require further work."

What it means
M&M's analysis clearly suggests that perhaps the most important data in the world today may not be as pristine as is needed to reveal the truth about what they are supposed to represent.  Someone needs to get on with the "further work" that is needed to resolve the issue.  And quickly.