In today’s age of fast and multifarious publications, we often come across such paradoxes, for example, that research and statistical analysis may indeed have been carried out correctly (properly), but the interpretation of the results is inadequate, or even incorrect, or misleading. To relate this hypothesis to Moravec’s paradox, one could reformulate the latter by saying that ‘to do research is easy; to discuss results is difficult’. But why is this so? Let us consider this issue from the perspective of another paradox, Simpson’s paradox.