Strategies for Improving Robustness in Experiments and Treatment Effects

Yasmin Valdez  | 27 de septiembre de 2018  | Vistas: 43

Raymond Duch is the director of the Nuffield Center for Experimental Social Sciences, also a researcher and professor. This conference is focused on the analysis of replicated data, the average treatment effect, cheating and strategies for improving robustness in Social Sciences investigation.

Replication of data and articles in different research disciplines is common today; this is what Duch names poor science. He presents a recent initiative in which he is involved to ensure that the average treatment effect (ATE) is more effective. He begins questioning if there's any reason to believe that the treatment effect is robust enough to be the basis for a policy recommendation. He comments that when he advises government entities, they do not have a solid basis for making political decisions.

Data fabrication, poor science its pervasive in the social sciences.”

Raymond suggests and explains three critical strategies to ensure robustness in the Social Sciences:

  • Diversity
  • Power
  • Internal Validity

The professor presents the Centre for Experimental Social Sciences, created to promote these three experimental analysis concerns; currently, there are centers in the United Kingdom, Chile, and India. He describes the consistent process of CESS to generate real data, efficient results, avoiding violations and fraud.

Cheating it’s not only amongst the academics but cheating it’s prevalent in our society, it’s costly.”

His most recent experiment is related to dealing with dishonesty, in his perspective is important to figure out who cheats in society. He concludes discussing the results of the method and discovers an association with abilities, performance, and success.


Director, Nuffield Centre for Experimental Social Sciences