Among the defining characteristics of a healthy research discipline is ability to correct its knowledge if more recent evidence creates grounds for this. Studies that reveal errors in earlier theories demonstrate, in line with Karl Popper's thinking, an approach called falsificationism. They complement approaches aimed at developing and expanding knowledge by generalising empirical observations or postulating new contributions and testing them. The paper presents analysis that applies this categorisation to abstracts of research papers (N = 5,202) in the eight leading IS journals. Machine-learning-based classification determined that only 7.0% of the papers manifested any clear form of knowledge-contestation, such as falsification, in the approach or findings presented. In light of this, we call on IS researchers to increase the falsification and knowledgecontestation in their research, to nurture more valid theories, methods, and practices, thereby achieving greater societal impact. We present two suitable IS research designs accordingly: knowledge-contesting comparisons and knowledge-contesting replications. We also discuss how these designs, exemplifying opportunities to increase the number of knowledge-contesting studies in the field, can be applied in both positivist and interpretivist research epistemology.