Research in humanities disciplines such as anthropology, archaeology, linguistics and theology can learn from replication failures in the biomedical and social sciences (go.nature.com/2stme7r).
Replication studies are not unprecedented in the humanities. The deciphering of Egyptian hieroglyphics was validated by comparing the Demotic, hieroglyphic and ancient Greek texts on the Rosetta stone found in 1799, for example. In 2013, the painting Sunset at Montmajour was confirmed as a genuine work by Vincent van Gogh after consulting letters by the artist describing it, and after analysing its chemical composition, colours and themes.
Replicability testing is particularly important for results in humanities disciplines that use empirical methods, and for cornerstone studies. Existing data sets can be reanalysed, or new data can be collected using the same or a modified study protocol (direct or conceptual replication, respectively). Conceptual replication is useful because it allows researchers to triangulate results.
Such testing will depend on preregistration of studies and on providing public access to detailed methods, data-analysis plans and data sets. It is also important to develop and use reporting guidelines for study protocols, publications and data sets. Funding agencies and scientific journals can help by demanding transparency and by funding and publishing replication studies.
Nature 558, 372 (2018)