Social scientific findings have been applied to resolve problems in many sectors of society, including higher education, the criminal justice system, traditional and social media, health systems, and politics. This would be excellent news if the social sciences had a track record of producing reliable findings that generate productive interventions. Unfortunately, many of the most prominent social scientific findings have proven to be unreliable. And where basic knowledge is unreliable, interventions based on that knowledge are likely to fail.

One contributing cause to this problem is that scholars have excessive freedom to select study materials and procedures optimally designed to confirm their favored hypotheses (rather than to expose their hypotheses to the most rigorous tests). Consequently, before research is even conducted, the methods themselves are frequently biased in favor of particular conclusions. Scientists are humans, and if they are not held accountable to Mertonian norms of impartiality and organized skepticism (Merton, 1942/1973), we should expect them to fall prey to the same types of cognitive biases and motivated reasoning as ordinary mortals.

Adversarial collaborations seek to raise the bar for scholars by increasing accountability at each stage of the research process, from articulating a fair research question, designing balanced methods, committing to belief updating depending on the results, and judiciously communicating those results. Putting scholars’ hypotheses and theories to the most rigorous tests incentivizes them to forward the versions of their arguments that are most defensible, rather than the most sensational. Moreover, the results of these adversarial collaborations have potential to minimize ambiguity in the literature because disagreeing scholars jointly publish results, forwarding one clearer (if more moderate and nuanced) state of the art, which helps other scientists have more accurate beliefs about empirical reality (enabling them to forward better hypotheses themselves).

The Adversarial Collaboration Project provides both instrumental and moral support to scholars with clashing theoretical-ideological views to work together in order to construct better public knowledge, more quickly and efficiently than traditional methods. Beyond advancing scientific debates and providing better information for educators, policymakers, organizations, and everyday people, we see the collegiality and transparency of adversarial collaborations as essential to re-establishing trust in the behavioral sciences.