The (forgotten) atomistic fallacy in Political Science and its implications for how we interpret elections

Improvements in the availability, accuracy and processing of individual level data have allowed political science literature to address the ‘ecological fallacy’, whereby inferences are made about individuals based on units of analyses operating at a higher level. Yet there has been limited attention to the risk that individual-level analyses may suffer from the reverse ‘atomistic’ – or ‘individualistic’- fallacy: the erroneous practice of drawing inferences about national level outcomes based on individual-level analyses. In this research note, we present a mathematical statement and simulations to diagnose and evaluate the extent of this fallacy in the case of voting behaviour. We also illustrate the problem using European Social Survey (ESS) data on far-right voting. We conclude by identifying three ‘perils’ of the atomistic fallacy, related to extrapolating conclusions about a party’s overall performance from information about an individual’s voting propensity. These perils can significantly affect how researchers interpret election results, and in turn, the policy implications of political science research.