- Joined
- Dec 22, 2012
- Messages
- 66,501
- Reaction score
- 22,160
- Location
- Portlandia
- Gender
- Male
- Political Leaning
- Libertarian - Right
From the most recent Nature Climate Change, editorial:
Perhaps the bigger question mark hovering over the act is what would qualify as replication. Many datasets are essentially not reproducible in practice: consider satellite data for a given period, or experiments over extended periods of time that are financially and logistically prohibitive to repeat. So perhaps only the data analysis must be replicated — but then to what degree of similarity must the findings correspond? Indeed, politically controversial findings that have repeatedly passed the test of replication, and moreover verification using alternative methods, have remained contested at a political level: the 'hockey stick' graph — showing how high recent temperatures are compared with the past 2,000 years — is perhaps an exemplar.
Political swings and roundabouts : Nature Climate Change : Nature Research
If datasets are not reproducible, what does it say about the quality of the data?
Perhaps the bigger question mark hovering over the act is what would qualify as replication. Many datasets are essentially not reproducible in practice: consider satellite data for a given period, or experiments over extended periods of time that are financially and logistically prohibitive to repeat. So perhaps only the data analysis must be replicated — but then to what degree of similarity must the findings correspond? Indeed, politically controversial findings that have repeatedly passed the test of replication, and moreover verification using alternative methods, have remained contested at a political level: the 'hockey stick' graph — showing how high recent temperatures are compared with the past 2,000 years — is perhaps an exemplar.
Political swings and roundabouts : Nature Climate Change : Nature Research
If datasets are not reproducible, what does it say about the quality of the data?