Is the Glass Half Empty or Half Full? Latest Results in the Replication Crisis in Psychology


For years a chorus of voices has spoken of a replication crisis in psychology, particularly social psychology. In November 2018, the results of another large replication project called Many Labs 2 were published, which demonstrated that in half of known psychological experiments it is possible to generate results that are similar to—but slightly lower than—the original results. Over a period of several years, 186 researchers from sixty different laboratories on six different continents representing thirty-six different nationalities conducted replications of twenty-eight well-known studies. Over 7,000 participants from diverse cultures took part in each of them. By comparison, the average number of participants in the original studies was 112. The results demonstrated clearly that replicability of experiments is not dependent on cultural context. Brian Nosek from the Center for Open Science, who helped to coordinate the study, claims that their results could prove even more significant than those of the 2015 Reproducibility Project, in which a mere 36 percent of experiments were successfully reproduced.

Studies were selected for replication based on a wide range of criteria, including frequency of citations, demonstrating the impact of a given study on the whole of psychology. Another, more technical criterion was the possibility to conduct the study online via the internet, using a platform accessible to researchers from all sixty laboratories.

Among the experiments whose results could not be confirmed was one demonstrating the Lady Macbeth effect: like the character from the Shakespearean drama, haunted by memories of her own crimes and convinced that she could see traces of blood on her hands, participants asked to recall some behavior associated with a violation of moral norms would have a strong tendency to engage in behaviors involving restoration and maintenance of physical cleanliness. Confirmation was also not achieved of results showing that people growing up with siblings are more altruistic than only children. There was no success in confirming the results of experiments showing that participants performing tasks involving the use of words describing heat are more likely to believe in global warming than those who performed tasks linked with the use of words describing cold. Replication was also not achieved of the famous 2009 experiment by Yoel Inbar showing that people who had first experienced a feeling of aversion were later more homophobic than those from the control group.

This article is available to subscribers only.
Subscribe now or log in to read this article.