Online / 6 & 7 February 2021


ReplicationWiki - Transparency in the Social Sciences

Informing about Data & Code Availability and Published Replications

The ReplicationWiki provides an overview of published empirical studies in the social sciences with information on data and code availability, data sources, and software. One can search for keywords, Journal of Economic Literature codes, and geographical origin of data. It informs about 670 replications, that is studies reanalyzing previously published results, as well as corrections and retractions. The wiki helps researchers to compare their results to those of previous studies. It is a resource that helps to identify useful teaching examples for statistical methods, replication and studies of social science. It allows advanced students and practicing researchers to find guidance on how to publish their replication research in various journals. A collection of teaching resources, useful tools, and literature helps instructors to integrate replication into their teaching and students to integrate open science practices into their own research. With the ongoing expansion of the wiki, currently covering more than 4,500 empirical studies, it is becoming an ever more powerful tool for social science research and education. It is a crowd-based platform where users can add their own replication results, suggest studies that should be replicated, and identify for example further data sources used in the empirical studies, especially ones from countries underrepresented in the literature and for whom economic policies are thus difficult to investigate.

The Wiki uses Semantic Media Wiki technology that is evolving, and a number of technical improvements in terms of usability and database structure are planned. A massive expansion of the content is planned based on machine learning and natural language processing techniques identifying the relevant information from the available data. For the technical improvements further expertise is welcome, and for the content expansion developers and researchers from all fields of the empirical social sciences are invited to join.


Jan H. Höffler