The aim of this paper is to provide a test of the Leys-Sartori conjecture. In its general version there is no support for their prediction that the smaller the district magnitude the higher the frequency of strategic voting at the primary district level using data of the first CSES data module. Nevertheless, after looking more closely at the particular seat-allocation rules in participating countries and distinguishing between SDSA (supra district seat-allocation) systems, where various situational criteria should undermine the Duvergerian logic, and LSA (local seat-allocation) systems, where these criteria are presumably absent, the main result of this paper shows that the predictions of the Leys-Sartori conjecture does hold for electoral districts in LSA systems but not for SDSA systems.