12 mars 2019
12h45-13h45
Louvain-la-Neuve
Leclercq 80
Philippe Bocquier (DEMO/UCLouvain),
Don't use the p-value! Alternative tests of significance
The p-value is probably the most popular significance test used in quantitative analysis. However, it is now increasingly questioned. Notwithstanding misinterpretation, p-hacking, and bias toward publishing positive results, what are the alternatives to the p-value for hypothesis testing? How can the collected data be used to produce a more robust significance test? This presentation will review the problems with p-values, present some theoretical alternatives, and propose an empirical alternative using survey data.