Replication: A View From the Streets

I support a replication policy because it would require political scientists to document carefully how they have generated their findings. Yes, creating a replication data set is a little extra work, but it is work researchers should be doing in the first place. Compensating for this cost are two benefits: The author is more sure that the results are valid, and the study has greater credibility simply because the author is willing to release the data and documentation to others. In two cases