It is concrete examples of incorrect treatment of concepts in comparative politics that are analyzed in the final part of the article (for the first two parts see Polis, 2003, №№ 3 and 4). With models examined, of the usage of the “structure” and “function” categories and of the “pluralism”, “integration”, “participation” and “mobilization” concepts as examples, the author demonstrates that the basic problem of comparative studies is confusion as to the level of analysis, which entails conceptual stretching and logical inconsistencies. As a result, the very purpose of comparing is defeated, and we are left to swim in a sea of empirical and theoretical messiness. To overcome this situation, according to Sartori, it is necessary and possible (1) to develop the discipline along a medium level of abstraction with better intermediate categories, and (2) to manoeuvre, both upwards and downwards, along a ladder of abstraction in such a way as to bring together assimilation and differentiation, a relatively high explanatory power and a relatively precise descriptive content, macro-theory and empirical testing.