|
Abstract:
|
Roberts et al. (2004) and Burnham et al. (2006) estimate excess mortality caused by the invasion and occupation of Iraq. Despite intensive media coverage and extensive discussion among statisticians about the technical details and merits of these studies, the authors refused to provide the underlying data and computer code that would allow outside researchers to confirm their results. How, then, can anyone know that those results are accurate? I argue that the authors should follow the replication standard, as outlined in King (1995), and provide their data and code, while maintaining respondent confidentiality. To encourage data sharing in other controversial settings, the American Statistical Association should promulgate the following Guideline: "The results of published studies which fail to meet the replication standard should be disregarded."
|
