Global warming caused by carbon emission has been recognized as a threat to public health and welfare. Carbon emission reduction is therefore a necessary task for each country to address the severe challenges arising from global warming. This research combines the STIRPAT model with the use of the panel and time-series data to analyze the impacts of population, affluence and technology on the carbon emission of 125 countries at different income levels over the period of 1990-2011. The results show that the key impact factor (KIF) at global level is affluence, followed by technology and population in the order of their impacts on carbon emission. For countries at high-income (HI) level, technology has the greatest impact on carbon emission, while affluence has the least. Affluence, prior to technology and population, is identified as the KIF of carbon emission for countries at upper-middle-income (UMI) and lower middle-income (LMI) levels. When it comes to the low-income (LI) level, affluence serves as the factor greatest affecting on carbon emission, and technology has the least impact. In particular, two generic patterns are identified based on the empirical results: higher income leads to greater impact of the technology and lower impact of the affluence on carbon emission. The KIFs of different income level countries identified in this study provide policy-makers and practitioners with valuable references for adopting effective policies and strategies to stimulate the global carbon emission reduction. (C) 2016 Elsevier Ltd. All rights reserved.