Typically, conflicts in world environmental negotiations are related, amongst other aspects, to the level of polarization of the countries in groups with conflicting interests. Given the predictable relationship between polarization and conflict, it would seem logical to evaluate the degree to which the distribution of countries – for example, in terms of their CO2 emissions per capita – would be structured through groups which in themselves are antagonistic, as well as their evolution over time. This paper takes the concept of polarization to explore this distribution for the period 1992-2010, looking at different analytic approaches related to the concept. Specifically, it makes a comparative evaluation of the results associated with endogenous multi-polarization measures (i.e. EGR and DER indices), exogenous measures (i.e. Z-K or multidimensional index) and strict bipolarization measures (i.e. Wolfson’s measure). Indeed, the interest lies not only in evaluating the global situation of polarization by comparing the different approaches and their temporal patterns, but also in examining the explanatory capacity of the different proxy groups used as a possible reference for designing global environmental policy from a group premise. JEL codes: D39; Q43; Q56. Key words: polarization; carbon emissions; conflict;