Thresholds of Corrections

In the article A Simple Excel calculation error puts a famous economic study under scrutiny by Nathan Ingrahamit research done by two economists Carmen Reinhard and Ken  Rogoff was found to have various errors, most notably an excel formula error and omitting some data in analysis. These errors where discovered by three University of Massachusetts researchers. Reinhard and Rogoff have acknowledged the excel errors, but defend that their analysis and conclusions are still valid. This could have been a shining example of the ability of the internet to correct highly sited studies that are found to be wrong. This correction could then be relayed out to the people that linked to the previous study and this dynamic ability of the internet could be fully realized in an academic setting, but it was not and the corrections to the excel formulas does not change the outcome all that much.
There is an interesting social network at play about cascading behavior. When the mistakes were first published, the original economists said that the UMass researchers where wrong at first, and then after some checking out the original study, the economists admitted to this mistake. One could say that Reinhard and Rogoff have a low threshold for accepting an easy to identify mistake, but the two disregards the claim that their omitting of data in their statistical analysis and the methods used to analyze their data skewed the results of the study or even happened. In this case, there are not enough people apparently agreeing with the UMass researchers for Reinhard and Rogoff to push them over their threshold in this case.
Seeing the motivations behind changing one’s position, behavior, or adopting A over B brings to bear the human elements involved in what can be explained as something very mathematical. We see the two economists stubbornness to admit their excel mistake become overcome as they are pressured into looking at the excel file again, but they stand strong on their findings despite the continued criticism. This could be in part of a bias to believe their own work superseding the possible network pressures being placed on them, but this should not be discounted when looking at cascading behavior.

Advertisements

Author: DonJavo

I am a content creator and co-founder of Where'd the Time Go, a company dedicated to videogame culture through filming lets plays and other videos. Don't hesitate to contact me here on WordPress if you have any questions or comments.