One cornerstone of modern science is its ability to self-correct. Scientific mistakes happen, but in a well-functioning science, they are supposed to be washed away by further experiments, which should correct past biases and flawed data. Science is pushed forward by mutual criticism; scientists are supposed to learn from their errors.

However, despite this general tendency towards improvement and self-correction, puzzling anomalies remain. One is the recent crisis in replicability, which has shown that many classic experiments in psychology and social science have failed to hold up upon scrutiny, even though they have been enshrined in textbooks for decades. Another source of puzzlement is the uneven progress of science, with some disciplines making major strides, and others being stuck with backward methods.

In a recent paper, Paul Smaldino and Cathy O’Connor (2020) have highlighted one major reason for this uneven progress of sciences: the lack of transmission of appropriate methods due to disciplinary insularity and parochialism. Backed by a wealth of evidence on peer-review practices, Smaldino and O’Connor argue that scientists tend to favour established methods within their respective disciplines, even if these methods are considered to be nefarious by several outsiders.

There are multiple reasons for this parochialism. First, scientists mostly read papers within their fields; this means that they may be ignorant of methods developed in other fields. Second, scientists have few incentives to be curious and try to learn methods from other fields. The principal goal for scientists is to publish, and publishing does not require developing the best methods. It simply requires promoting methods that are considered “good enough” to pass the scrutiny of peer-review, even though better methods could be available elsewhere. If the reviewers are not finding faults with the proposed methods, then it is of little use to try to propose improvements whose absence would have been unnoticed. This means that authors have an incentive to remain within the boundaries of what can be understood by the reviewers.

This familiarity with established methods can make scientists wary of any foreign methods, established in other fields. Due to lack of statistical training, scientists are often led to accept methods on blind faith, without having any clear idea of the validity of deep reasons behind their promotion. Two recent examples of scientific fields accepting deeply flawed methods, despite their immediate rejection by statisticians, involve Magnitude Based Inference in sport science, and the popularity of p-rep in psychology in the late 2000s (Iverson et al., 2009; Aisbett, 2020).

How can we escape the obstacles to scientific progress caused by entrenched parochial preferences? Smaldino and O’Connor highlight one natural solution: the promotion of interdisciplinary research. Researchers who cross borders tend to be less wedded to a particular method, and are thus well-positioned to be agents in the promotion of new and better methods.

This boost associated with interdisciplinary research can be illustrated by one study in which I recently participated. Our team, led by psychologist Liz Tenney (Tenney et al., 2020), studied the evolution of sample size and open science practices at four management journals during the period 2011-2019. Open science practices (Open data, Open materials, and the use of pre-registration) have been promoted by the recent credibility revolution in psychology, and have quickly become prominent in some psychology journals, such as the prestigious Psychological Science. In management, these practices have remained rare until 2019. However, one of the four management journals, OBHDP, did much better than the other journals and saw a strong increase in open science practices over the last years. Interestingly, OBHDP differs from the other journals by being interdisciplinary, crossing both management and psychology research. While this is just one example, it illustrates one interesting case of the spread of good research practices across disciplines.

One major aspect of scientific integrity involves adopting the best methods to promote scientific progress. In reverse, scientific integrity involves uprooting traditional biases and suboptimal research methods in order to limit the risk of incorrect inferences. Scientific integrity requires curiosity and interdisciplinary research.