Global, neat-surface temperature data sets and their derivations are discussed, and differences between the Jones and Intergovernmental Panel on Climate Change data sets are explained. Global-mean temperature changes are then interpreted in terms of anthropogenic forcing influences and natural variability: The inclusion of aerosol forcing improves the fit between modeled and observed changes hot does not improve the agreement between the implied climate sensitivity value and the standard model-based range of 1.5-4.5 degrees C equilibrium warming for a CO2 doubling, The implied sensitivity goes from below the model-based range of estimates to substantially above this range, The addition of a solar forcing effect. further improves the fit and brings the best-fit sensitivity into the middle of the model-based range, Consistency is further improved when internally generated changes are considered, This consistency, however, hides manly uncertainties that surround observed data/model comparisons, These uncertainties make it impossible currently to use observed global-scale temperature changes to narrow the uncertainty range in the climate sensitivity below that estimated directly from climate models.